Extract_User_Characteristics.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# This notebook studies the characteristics of the counterspeech users and Hate users."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import the required Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:33.766409Z",
     "start_time": "2018-11-02T18:47:33.753599Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "IPython.notebook.set_autosave_interval(5000)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Autosaving every 5 seconds\n"
     ]
    }
   ],
   "source": [
    "%autosave 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:35.100319Z",
     "start_time": "2018-11-02T18:47:33.831613Z"
    }
   },
   "outputs": [],
   "source": [
    "import json\n",
    "import pickle\n",
    "import os\n",
    "import re\n",
    "from collections import defaultdict,Counter,OrderedDict\n",
    "import gzip\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from scipy.stats import ttest_ind\n",
    "from scipy.stats import mannwhitneyu\n",
    "from scipy.stats import normaltest\n",
    "import scipy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analyzing our own dataset of 1513 tweets:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:35.160494Z",
     "start_time": "2018-11-02T18:47:35.102508Z"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import re\n",
    "df = pd.read_json('Refined_Project_Dataset/Binny_Tagged_Final.json')\n",
    "# df.drop_duplicates(subset = ['original_tweetid','reply_id'],inplace = True)\n",
    "# df.to_json('Refined_Project_Dataset/Binny_Tagged_Final.json',orient = 'records')\n",
    "df.head(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:35.242578Z",
     "start_time": "2018-11-02T18:47:35.162258Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1513, 10)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "223"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(df.shape)\n",
    "(df.tag=='0').sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Unique Hate and Counter Usernames:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:35.367240Z",
     "start_time": "2018-11-02T18:47:35.244341Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(548, 1239, 220)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hate_usernames = set(df['original_tweetusername'].values.tolist())\n",
    "counter_usernames = set(df[df.tag!='0']['replier_user_name'].values.tolist())\n",
    "Non_counter_usernames = set(df[df.tag=='0']['replier_user_name'].values.tolist())\n",
    "len(hate_usernames),len(counter_usernames),len(Non_counter_usernames)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Unique Hate and Counter UserIDs:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:35.500255Z",
     "start_time": "2018-11-02T18:47:35.369988Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(558, 1290, 223)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hate_usernamesID = set(df['original_tweetid'].values.astype(str))\n",
    "counter_usernamesID = set(df[df.tag!='0']['reply_id'].values.astype(str))\n",
    "Non_counter_usernamesID = set(df[df.tag=='0']['reply_id'].values.astype(str))\n",
    "len(hate_usernamesID),len(counter_usernamesID),len(Non_counter_usernamesID)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Deleted  users:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:35.687591Z",
     "start_time": "2018-11-02T18:47:35.502254Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "85"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "deleted_tweets = np.loadtxt('deleted/Deleted_tweets.txt',dtype = str,delimiter = '\\n')\n",
    "deleted_users = np.loadtxt('deleted/Deleted_users.txt',dtype = str,delimiter = '\\n')\n",
    "users = [] \n",
    "for user in deleted_users:\n",
    "    if '404' in user:\n",
    "        users.append(re.findall(r'\\w+',user)[0])\n",
    "len(users)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:35.751484Z",
     "start_time": "2018-11-02T18:47:35.689288Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(25, 53, 8)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Deleted_hate = list(set(users).intersection(set(hate_usernames)))\n",
    "Deleted_counter = list(set(users).intersection(set(counter_usernames)))\n",
    "Deleted_Noncounter = list(set(users).intersection(set(Non_counter_usernames)))\n",
    "print(len((hate_usernames.intersection(counter_usernames)).intersection(Deleted_counter)))\n",
    "len(Deleted_hate),len(Deleted_counter),len(Deleted_Noncounter)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:35.898857Z",
     "start_time": "2018-11-02T18:47:35.753159Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Noncounter :  {'0': 8, '1': 0, '2': 0, '3': 0, '4': 0, '5': 0, '6': 0, '7': 0, '8': 0, '9': 0, '11': 0}\n",
      "\n",
      "Counter :  {'0': 0, '1': 6, '2': 3, '3': 6, '4': 1, '5': 4, '6': 15, '7': 4, '8': 10, '9': 10, '11': 7}\n",
      "\n",
      "Hate :  {'0': 2, '1': 0, '2': 2, '3': 2, '4': 0, '5': 3, '6': 1, '7': 4, '8': 5, '9': 9, '11': 3}\n"
     ]
    }
   ],
   "source": [
    "missing_counter_labels = []\n",
    "missing_hate_labels = []\n",
    "missing_Noncounter_labels = []\n",
    "for i in counter_usernames:\n",
    "    if i in Deleted_counter:\n",
    "        missing_counter_labels.append(df[df['replier_user_name']== i]['tag'].values[0])\n",
    "\n",
    "for i in hate_usernames:\n",
    "    if i in Deleted_hate:\n",
    "        missing_hate_labels.append(df[df['original_tweetusername']== i]['tag'].values[0])\n",
    "\n",
    "for i in Non_counter_usernames:\n",
    "    if i in Deleted_Noncounter:\n",
    "        missing_Noncounter_labels.append(df[df['replier_user_name']== i]['tag'].values[0])\n",
    "# np.savetxt('missing_counter_labels.txt',missing_counter_labels,fmt = '%s') \n",
    "# np.savetxt('missing_hate_labels.txt',missing_hate_labels,fmt = '%s') \n",
    "\n",
    "categories = ['0','1','2','3','4','5','6','7','8','9','11']\n",
    "dict_counter = {}\n",
    "dict_hate = {}\n",
    "dict_Noncounter = {}\n",
    "for i in categories:\n",
    "    dict_counter[i] = (re.findall(r'\\d+',(' '.join(missing_counter_labels)))).count(i)\n",
    "                \n",
    "for i in categories:\n",
    "    dict_hate[i] = (re.findall(r'\\d+',(' '.join(missing_hate_labels)))).count(i)\n",
    "for i in categories:\n",
    "    dict_Noncounter[i] = (re.findall(r'\\d+',(' '.join(missing_Noncounter_labels)))).count(i)\n",
    "    \n",
    "print('Noncounter : ',dict_Noncounter,end = '\\n\\n')\n",
    "print('Counter : ',dict_counter,end = '\\n\\n')\n",
    "print('Hate : ',dict_hate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.018660Z",
     "start_time": "2018-11-02T18:47:35.903125Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1972"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#######UserNames of all Counter,non-counter and Hate Users:\n",
    "all_usernames = [i for i in (hate_usernames.union(counter_usernames)).union(Non_counter_usernames)]\n",
    "len(all_usernames)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.191511Z",
     "start_time": "2018-11-02T18:47:36.020849Z"
    }
   },
   "outputs": [],
   "source": [
    "h,c,n,a = [],[],[],[]\n",
    "deleted = np.loadtxt('Deleted_tweets.txt',delimiter = '\\n' , dtype = str)\n",
    "for each_id in deleted:\n",
    "    if re.findall(r'\\w+',each_id)[-1]== 'USER_DELETED':\n",
    "        print(end = '')\n",
    "#         if re.findall(r'\\w+',each_id)[0] in hate_usernames:\n",
    "#             h.append(re.findall(r'\\w+',each_id)[0])\n",
    "#         elif re.findall(r'\\w+',each_id)[0] in counter_usernames:\n",
    "#             c.append(re.findall(r'\\w+',each_id)[0])\n",
    "#         elif re.findall(r'\\w+',each_id)[0] in Non_counter_usernames:\n",
    "#             n.append(re.findall(r'\\w+',each_id)[0])\n",
    "    else:\n",
    "        a.append(re.findall(r'\\w+',each_id)[0])\n",
    "    if each_id in hate_usernamesID:\n",
    "        h.append(each_id)\n",
    "    elif each_id in counter_usernamesID:\n",
    "        c.append(each_id)\n",
    "    elif each_id in Non_counter_usernamesID:\n",
    "        n.append(each_id)\n",
    "hca = []\n",
    "for i in a:\n",
    "    if i in hate_usernames:\n",
    "        hca.append(df[df['original_tweetusername']==i]['original_tweetid'].values[0])\n",
    "    if i in counter_usernames.union(Non_counter_usernames):\n",
    "        hca.append(df[df['replier_user_name']==i]['reply_id'].values[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.294948Z",
     "start_time": "2018-11-02T18:47:36.194175Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(169, 412, 84)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# (set(h).intersection(set(c)).union(set(h).intersection(set(n)))).union(set(c).intersection(set(n)))\n",
    "len(set(h)^(set(h).intersection((set(c).union(set(n)).union(hca))))) , len(set(c)^(set(c).intersection((set(n).union(set(h)).union(hca))))) , len(set(n)^(set(n).intersection((set(c).union(set(h)).union(hca)))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.402737Z",
     "start_time": "2018-11-02T18:47:36.297537Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(169, 412, 84)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "h = set(h)-(set(h).intersection(hca))\n",
    "len(set(h)),len(c),len(n)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.509640Z",
     "start_time": "2018-11-02T18:47:36.405532Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "169"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(set(h)-(set(h).intersection(hca)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Finding out intersections b/w users :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.608556Z",
     "start_time": "2018-11-02T18:47:36.512380Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(len(hate_usernames.intersection(counter_usernames)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.716961Z",
     "start_time": "2018-11-02T18:47:36.610503Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(hate_usernames.intersection(Non_counter_usernames))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.818048Z",
     "start_time": "2018-11-02T18:47:36.719374Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(counter_usernames.intersection(Non_counter_usernames))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Category Wise Distribution:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:36.932464Z",
     "start_time": "2018-11-02T18:47:36.822638Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Category0 Count:223\n",
      "Category1 Count:136\n",
      "Category2 Count:177\n",
      "Category3 Count:69\n",
      "Category4 Count:58\n",
      "Category5 Count:132\n",
      "Category6 Count:168\n",
      "Category7 Count:134\n",
      "Category8 Count:175\n",
      "Category9 Count:357\n",
      "Category11 Count:198\n"
     ]
    }
   ],
   "source": [
    "categories = ['0','1','2','3','4','5','6','7','8','9','11']\n",
    "for i in categories:   \n",
    "    print('Category{} Count:{}'.format(i,(re.findall(r'\\d+',(' '.join(df.tag.tolist())))).count(i)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pic Based analysis:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:46.225811Z",
     "start_time": "2018-11-02T18:47:36.935440Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12, 145, 60)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "########### For Hate ###########\n",
    "def get_hateids_with_pics(names):\n",
    "    ids = []\n",
    "    for name in names:\n",
    "        tweets = df[df['original_tweetusername']==name].drop_duplicates(subset = ['original_tweetid'])['original_tweet'].tolist()\n",
    "        IDs = df[df['original_tweetusername']==name].drop_duplicates(subset = ['original_tweetid'])['original_tweetid'].tolist()\n",
    "        if len(tweets)!=1:\n",
    "            for tweet,ID in zip(tweets,IDs):\n",
    "                if re.findall(r'pic.twitter.com/',tweet)!= list():\n",
    "                    ids.append(ID)\n",
    "        else:\n",
    "            if re.findall(r'pic.twitter.com/',tweets[0])!= list():\n",
    "                    ids.append(IDs[0])           \n",
    "    return ids\n",
    "########### For Non-Counter ###########\n",
    "def get_Noncounterids_with_pics(names):\n",
    "    ids = []\n",
    "    for name in names:\n",
    "        tweets = df[df['replier_user_name']==name].drop_duplicates(subset = ['reply_id'])['reply_text'].tolist()\n",
    "        IDs = df[df['replier_user_name']==name].drop_duplicates(subset = ['reply_id'])['reply_id'].tolist()\n",
    "        if len(tweets)!=1:\n",
    "            for tweet,ID in zip(tweets,IDs):\n",
    "                if re.findall(r'pic.twitter.com/',tweet)!= list():\n",
    "                    ids.append(ID)\n",
    "        else:\n",
    "            if re.findall(r'pic.twitter.com/',tweets[0])!= list():\n",
    "#                     print(IDs[0],tweets)\n",
    "                    ids.append(IDs[0])           \n",
    "    return ids\n",
    "######### ID Based Analysis ##############\n",
    "hate_ids_pic = get_hateids_with_pics(hate_usernames)\n",
    "counter_ids_pic = get_Noncounterids_with_pics(counter_usernames) #### For Counter\n",
    "NonCounter_ids_pic = get_Noncounterids_with_pics(Non_counter_usernames)\n",
    "len(hate_ids_pic),len(counter_ids_pic),len(NonCounter_ids_pic)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:46.230646Z",
     "start_time": "2018-11-02T18:47:46.227943Z"
    }
   },
   "outputs": [],
   "source": [
    "# df.to_json('Refined_Project_Dataset/Binny_Tagged_Final.json',orient = 'records')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Let's see an example of hate users using pics:|"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-06T00:43:31.398997Z",
     "start_time": "2018-11-06T00:43:31.309051Z"
    }
   },
   "outputs": [],
   "source": [
    "df[df['original_tweetusername']=='T_Bowen10']['original_tweet'].values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pic to Pic Conversations:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:46.443153Z",
     "start_time": "2018-11-02T18:47:46.310944Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(69, 12)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "counter_pics = 0\n",
    "for i in hate_ids_pic:\n",
    "    new_df = df[df['original_tweetid']==i]\n",
    "    for j in counter_ids_pic:\n",
    "#         print(new_df['reply_id'].tolist().count(j))\n",
    "        counter_pics+=new_df['reply_id'].tolist().count(j)\n",
    "\n",
    "Noncounter_pics = 0\n",
    "for i in hate_ids_pic:\n",
    "    new_df = df[df['original_tweetid']==i]\n",
    "    for j in NonCounter_ids_pic:\n",
    "#         print(new_df['reply_id'].tolist().count(j))\n",
    "        Noncounter_pics+=new_df['reply_id'].tolist().count(j)\n",
    "counter_pics,Noncounter_pics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:46.584503Z",
     "start_time": "2018-11-02T18:47:46.445849Z"
    }
   },
   "outputs": [],
   "source": [
    "# for Id in counter_ids_pic:\n",
    "# #     print()\n",
    "#     if re.findall(r'pic.twitter.com/',df[df['reply_id']==Id]['reply_text'].values[0])!= list(): \n",
    "#         if df[df['reply_id']==Id]['tag'].values[0]!= '6':\n",
    "#             if df[df['reply_id']==Id]['tag'].values[0]!= '11':\n",
    "#                 print(df.loc[df['reply_id']==Id ,'tag'].values[0],end = '--------')\n",
    "#                 df.loc[df['reply_id']==Id ,'tag'] +=',6'\n",
    "#                 print(df.loc[df['reply_id']==Id ,'tag'].values[0])\n",
    "#             else:\n",
    "#                 print(df.loc[df['reply_id']==Id ,'tag'].values[0],end = '--------')\n",
    "#                 df.loc[df['reply_id']==Id ,'tag'] = '6'\n",
    "#                 print(df.loc[df['reply_id']==Id ,'tag'].values[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#######################################__End__######################################"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## User-Characteristics Analysis :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:46.947072Z",
     "start_time": "2018-11-02T18:47:46.586395Z"
    }
   },
   "outputs": [],
   "source": [
    "from empath import Empath\n",
    "from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer\n",
    "lexicon = Empath()\n",
    "sentiment_analyzer = SentimentIntensityAnalyzer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:47.024281Z",
     "start_time": "2018-11-02T18:47:46.949721Z"
    }
   },
   "outputs": [],
   "source": [
    "with open('Refined_Project_Dataset/lexicon_hate_user.pkl','rb') as fp:\n",
    "    lexicon_hate_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/lexicon_counter_user.pkl','rb') as fp:\n",
    "    lexicon_counter_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/lexicon_Noncounter_user.pkl','rb') as fp:\n",
    "    lexicon_Noncounter_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/sentiment_hate_user.pkl','rb') as fp:\n",
    "    sentiment_hate_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/sentiment_counter_user.pkl','rb') as fp:\n",
    "    sentiment_counter_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/sentiment_Noncounter_user.pkl','rb') as fp:\n",
    "    sentiment_Noncounter_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/textblob_hate_user.pkl','rb') as fp:\n",
    "    textblob_hate_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/textblob_counter_user.pkl','rb') as fp:\n",
    "    textblob_counter_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/textblob_Noncounter_user.pkl','rb') as fp:\n",
    "    textblob_Noncounter_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/profane_hate_user.pkl','rb') as fp:\n",
    "    profane_hate_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/profane_counter_user.pkl','rb') as fp:\n",
    "    profane_counter_user = pickle.load(fp)\n",
    "with open('Refined_Project_Dataset/profane_Noncounter_user.pkl','rb') as fp:\n",
    "    profane_Noncounter_user = pickle.load(fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.317061Z",
     "start_time": "2018-11-02T18:47:47.026686Z"
    }
   },
   "outputs": [],
   "source": [
    "import csv\n",
    "Hate_hate = []\n",
    "counter_hate = []\n",
    "Noncounter_hate = []\n",
    "c = 0\n",
    "with open('Refined_Project_Dataset/user_predictions_hate.csv') as fp:\n",
    "    for user_dict in csv.reader(fp,delimiter = '\\n'):\n",
    "        name = re.findall(r'\\w+',user_dict[0])[0]\n",
    "#         print(name)\n",
    "        lst = [int(i) for i in re.findall(r'\\d+',user_dict[0])]\n",
    "        c +=1\n",
    "        if name in hate_usernames:\n",
    "            Hate_hate.append(lst.count(0)/len(lst))\n",
    "        if name in counter_usernames:\n",
    "            counter_hate.append(lst.count(0)/len(lst))\n",
    "        if name in Non_counter_usernames:\n",
    "            Noncounter_hate.append(lst.count(0)/len(lst))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.317826Z",
     "start_time": "2018-11-02T13:11:44.215Z"
    }
   },
   "outputs": [],
   "source": [
    "#Hate\n",
    "print('Hate_Mean {}'.format(np.mean(Hate_hate)))\n",
    "print('Counter_Mean {}'.format(np.mean(counter_hate)))\n",
    "print('Noncounter_Mean {}'.format(np.mean(Noncounter_hate)))\n",
    "print('Hate_Median {}'.format(np.median(Hate_hate)))\n",
    "print('Counter_Median {}'.format(np.median(counter_hate)))\n",
    "print('Noncounter_Median {}'.format(np.median(Noncounter_hate)))\n",
    "####### Test :\n",
    "print('Ttest \\n Counter----Hate \\n',ttest_ind(counter_hate,Hate_hate))\n",
    "print('Counter----Noncounter \\n',ttest_ind(counter_hate,Noncounter_hate))\n",
    "print('Hate----Noncounter \\n',ttest_ind(Hate_hate,Noncounter_hate))\n",
    "print('Manwhitneytest \\n Counter----Hate \\n',mannwhitneyu(counter_hate,Hate_hate))\n",
    "print('Counter----Noncounter \\n',mannwhitneyu(counter_hate,Noncounter_hate))\n",
    "print('Hate----Noncounter \\n',mannwhitneyu(Hate_hate,Noncounter_hate))\n",
    "######################## Profane Hate ############################\n",
    "Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(Noncounter_hate))],Noncounter_hate],columns = ['Users',''])\n",
    "counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(counter_hate))],counter_hate],columns = ['Users',''])\n",
    "hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(Hate_hate))],Hate_hate],columns = ['Users',''])\n",
    "Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "plt.figure(figsize = (2.2,2))\n",
    "ax  = sns.boxplot(x = 'Users',y ='',data = Comb_df, linewidth=1,width=0.6,showfliers = False,palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "ax.set_xticklabels('')\n",
    "ax.set_xlabel('')\n",
    "ax.set_title('Hate')\n",
    "plt.tight_layout()\n",
    "plt.savefig('Images/Hate.eps', format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.318732Z",
     "start_time": "2018-11-02T13:11:44.297Z"
    }
   },
   "outputs": [],
   "source": [
    "#Offensive\n",
    "print('Hate_Mean {}'.format(np.mean(Hate_hate)))\n",
    "print('Counter_Mean {}'.format(np.mean(counter_hate)))\n",
    "print('Noncounter_Mean {}'.format(np.mean(Noncounter_hate)))\n",
    "print('Hate_Median {}'.format(np.median(Hate_hate)))\n",
    "print('Counter_Median {}'.format(np.median(counter_hate)))\n",
    "print('Noncounter_Median {}'.format(np.median(Noncounter_hate)))\n",
    "####### Test :\n",
    "print('Ttest \\n Counter----Hate \\n',ttest_ind(counter_hate,Hate_hate))\n",
    "print('Counter----Noncounter \\n',ttest_ind(counter_hate,Noncounter_hate))\n",
    "print('Hate----Noncounter \\n',ttest_ind(Hate_hate,Noncounter_hate))\n",
    "print('Manwhitneytest \\n Counter----Hate \\n',mannwhitneyu(counter_hate,Hate_hate))\n",
    "print('Counter----Noncounter \\n',mannwhitneyu(counter_hate,Noncounter_hate))\n",
    "print('Hate----Noncounter \\n',mannwhitneyu(Hate_hate,Noncounter_hate))\n",
    "######################## Profane Offensive ############################\n",
    "Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(Noncounter_hate))],Noncounter_hate],columns = ['Users',''])\n",
    "counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(counter_hate))],counter_hate],columns = ['Users',''])\n",
    "hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(Hate_hate))],Hate_hate],columns = ['Users',''])\n",
    "Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "plt.figure(figsize = (2.2,2))\n",
    "ax  = sns.boxplot(x = 'Users',y ='',data = Comb_df, linewidth=1,width=0.6,showfliers = False,palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "ax.set_xticklabels('')\n",
    "ax.set_xlabel('')\n",
    "ax.set_title('Offensive')\n",
    "plt.tight_layout()\n",
    "plt.savefig('Images/Offensive.eps', format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Counterspeech Category Based Analysis:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.319562Z",
     "start_time": "2018-11-02T13:11:44.379Z"
    }
   },
   "outputs": [],
   "source": [
    "df[df['tag']=='1,5,9']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.320456Z",
     "start_time": "2018-11-02T13:11:44.468Z"
    }
   },
   "outputs": [],
   "source": [
    "a,c,c1,c2 = 0,0,0,0\n",
    "for i in df[df['tag']!='0']['tag'].tolist():\n",
    "    if len(re.findall(r'\\d+',i))==2:\n",
    "        c1+=1\n",
    "#         print(i)\n",
    "    elif len(re.findall(r'\\d+',i))==3 :\n",
    "#         print(i)\n",
    "        c2+=1\n",
    "    elif len(re.findall(r'\\d+',i))==1 :\n",
    "#         print(i)\n",
    "        c+=1\n",
    "    else:\n",
    "        print(i)\n",
    "        a+=1\n",
    "print(\"\\n1 Label :\",c,\"\\n2 Labels :\",c1,\"\\n3 labels:\",c2,\"\\n4 labels: 2\",\"\\n5 labels: 1\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.321257Z",
     "start_time": "2018-11-02T13:11:44.539Z"
    }
   },
   "outputs": [],
   "source": [
    "When_acting_as_CounterSpeakers_categories_in_which_they_have_tweeted_against_other_hate_speeches =[]\n",
    "When_acting_as_HateSpeakers_categories_in_which_their_hatespeech_has_been_countered_by_other_counter_speakers = []\n",
    "CounterSpeakers_categories = []\n",
    "HateSpeakers_categories = []\n",
    "for name in hate_usernames.intersection(counter_usernames):\n",
    "        CounterSpeakers_categories.append(df[df['replier_user_name']== name]['tag'].tolist())\n",
    "        HateSpeakers_categories.append(df[df['original_tweetusername']== name]['tag'].tolist())\n",
    "# for ind_name in hate_usernames.intersection(counter_usernames):   \n",
    "#     if name == ind_name:\n",
    "#         categories_in_which_their_hatespeech_has_been_tweeted_by_other_counter_speakers.append(tweets[tweets['original_tweetusername']== name]['tag'].values[0])\n",
    "for i in CounterSpeakers_categories:\n",
    "    When_acting_as_CounterSpeakers_categories_in_which_they_have_tweeted_against_other_hate_speeches += i \n",
    "for i in HateSpeakers_categories:\n",
    "    When_acting_as_HateSpeakers_categories_in_which_their_hatespeech_has_been_countered_by_other_counter_speakers += i "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.322054Z",
     "start_time": "2018-11-02T13:11:44.595Z"
    }
   },
   "outputs": [],
   "source": [
    "df.head(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.322798Z",
     "start_time": "2018-11-02T13:11:44.653Z"
    }
   },
   "outputs": [],
   "source": [
    "CounterSpeakers_categories,len(CounterSpeakers_categories)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.323674Z",
     "start_time": "2018-11-02T13:11:44.716Z"
    }
   },
   "outputs": [],
   "source": [
    "categories = ['0','1','2','3','4','5','6','7','8','9','11']\n",
    "dict_counter = {}\n",
    "dict_hate = {}\n",
    "for i in categories:\n",
    "    dict_counter[i] = (re.findall(r'\\d+',(' '.join(When_acting_as_CounterSpeakers_categories_in_which_they_have_tweeted_against_other_hate_speeches)))).count(i)\n",
    "                \n",
    "for i in categories:\n",
    "    dict_hate[i] = (re.findall(r'\\d+',(' '.join(When_acting_as_HateSpeakers_categories_in_which_their_hatespeech_has_been_countered_by_other_counter_speakers)))).count(i)\n",
    "print(dict_counter)\n",
    "print(dict_hate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.324736Z",
     "start_time": "2018-11-02T13:11:44.776Z"
    }
   },
   "outputs": [],
   "source": [
    "When_acting_as_HateSpeakers_categories_in_which_their_hatespeech_has_been_countered_by_other_counter_speakers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Since the file is too large to process, we will break it into multiple smaller pickle file for efficiency reasons."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.325545Z",
     "start_time": "2018-11-02T13:11:44.882Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# dict_user_info = {}\n",
    "# dict_user_tweetid = defaultdict(set)\n",
    "# dict_tweet_info = {}\n",
    "# set_tweet_info = set()\n",
    "# ##### Opening Tweet_Info File:\n",
    "# tweet_info_fp = open('Refined_Project_Dataset/tweet_info.json','w')\n",
    "\n",
    "# with open('Refined_Project_Dataset/Tweet_and_User_Info.json','r') as fp:\n",
    "#     for line_counter, line in enumerate(fp):\n",
    "#         if line_counter%100000==0:\n",
    "#             print(line_counter)\n",
    "#         temp_user = json.loads(line)['user']\n",
    "#         temp_tweet = json.loads(line)\n",
    "# #         print(type(temp_tweet))\n",
    "#         if temp_user['name'] not in dict_user_info:\n",
    "#             dict_user_info[temp_user['screen_name']] = temp_user\n",
    "        \n",
    "#         if temp_tweet['id_str'] not in set_tweet_info:\n",
    "# #             dict_tweet_info[temp_tweet['id_str']] = {'tweet_id':temp_tweet['id_str'], 'tweet_text': temp_tweet['text'],\n",
    "# #                                                     'created_at':temp_tweet['created_at'], 'entities': temp_tweet['entities'],\n",
    "# #                                                     'username': temp_tweet['user']['name']}\n",
    "#             set_tweet_info.add(temp_tweet['id_str'])\n",
    "#         if temp_tweet['retweeted']== False:\n",
    "#             temp_tweet_str = json.dumps({'tweet_id':temp_tweet['id_str'], 'tweet_text': temp_tweet['text'],\n",
    "#                                                     'created_at':temp_tweet['created_at'], 'entities': temp_tweet['entities'],\n",
    "#                                                     'username': temp_tweet['user']['screen_name']})\n",
    "#             tweet_info_fp.write(temp_tweet_str)\n",
    "#             tweet_info_fp.write('\\n')\n",
    "    \n",
    "# #         print(temp_user['name'], temp_tweet['id_str'])\n",
    "#         dict_user_tweetid[temp_user['screen_name']].add(temp_tweet['id_str'])\n",
    "        \n",
    "\n",
    "# with open('Refined_Project_Dataset/user_info.pkl','wb') as fp:\n",
    "#     pickle.dump(dict_user_info,fp)\n",
    "# print('User info pickle done..')\n",
    "\n",
    "# with open('Refined_Project_Dataset/user_tweetid.pkl','wb') as fp:\n",
    "#     pickle.dump(dict_user_tweetid,fp)\n",
    "# print('User tweetid mapping done...')\n",
    "\n",
    "# # with open('./Dataset/tweet_info.pkl','wb') as fp:\n",
    "# #     pickle.dump(dict_tweet_info,fp)\n",
    "# tweet_info_fp.close()\n",
    "# print('Tweet info pickle done...')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### tweet_info.json : Contains information about each id and its corresponding TweetData except user information. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.326403Z",
     "start_time": "2018-11-02T13:11:44.977Z"
    }
   },
   "outputs": [],
   "source": [
    "# all_user_tweets = defaultdict(list)\n",
    "# binny_hashtags =  {}\n",
    "# binny_hashtags['hate'],binny_hashtags['counter'],binny_hashtags['Noncounter'] = defaultdict(list),defaultdict(list),defaultdict(list)\n",
    "# with open('Refined_Project_Dataset/tweet_info.json','r') as fp:\n",
    "#     for line_counter,line in enumerate(fp):\n",
    "#         temp = json.loads(line.strip())\n",
    "# #         print(temp)\n",
    "# #         break\n",
    "#         print(line_counter)\n",
    "#         if temp['entities']['hashtags']!=[]:\n",
    "#             if temp['username'] in hate_usernames:\n",
    "#                 binny_hashtags['hate'][temp['username']].append(temp['entities']['hashtags'][0]['text'])\n",
    "#             elif temp['username'] in counter_usernames:\n",
    "#                 binny_hashtags['counter'][temp['username']].append(temp['entities']['hashtags'][0]['text'])\n",
    "#             elif temp['username'] in Non_counter_usernames:\n",
    "#                 binny_hashtags['Noncounter'][temp['username']].append(temp['entities']['hashtags'][0]['text'])\n",
    "# #         all_user_tweets[temp['username']].append(temp['tweet_text'])\n",
    "# # with open('Refined_Project_Dataset/all_user_tweets.pkl','wb') as fp:\n",
    "# #     pickle.dump(all_user_tweets, fp)\n",
    "with open('Refined_Project_Dataset/all_user_tweets.pkl','rb') as fp:\n",
    "    all_user_tweets = pickle.load(fp)\n",
    "# with open('Refined_Project_Dataset/Binny_hashtag_words_per_user.pkl','wb') as fp:\n",
    "#     pickle.dump(binny_hashtags,fp)\n",
    "# len(binny_hashtags['hate'].keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.327434Z",
     "start_time": "2018-11-02T13:11:45.035Z"
    }
   },
   "outputs": [],
   "source": [
    "# len(binny_hashtags['hate'].keys()),len(binny_hashtags['counter'].keys()),len(binny_hashtags['Noncounter'].keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-02T18:47:51.328329Z",
     "start_time": "2018-11-02T13:11:45.089Z"
    }
   },
   "outputs": [],
   "source": [
    "len(list(all_user_tweets.values())[1532])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Ravina Dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:43:37.607592Z",
     "start_time": "2018-10-29T01:43:37.600820Z"
    }
   },
   "outputs": [],
   "source": [
    "# all_tweets_corresponding_to_hate_usernames = {}\n",
    "# all_tweets_corresponding_to_counter_usernames = {}\n",
    "# for name,tweets in all_user_tweets.items():\n",
    "#     if name in hate :\n",
    "#         all_tweets_corresponding_to_hate_usernames.update({name:tweets})\n",
    "#     elif name in counter:\n",
    "#         all_tweets_corresponding_to_counter_usernames.update({name:tweets})\n",
    "\n",
    "# with open('all_tweets_corresponding_to_counter_usernames.pkl','wb') as fp :\n",
    "#     pickle.dump(all_tweets_corresponding_to_counter_usernames,fp)\n",
    "# with open('all_tweets_corresponding_to_hate_usernames.pkl','wb') as fp :\n",
    "#     pickle.dump(all_tweets_corresponding_to_hate_usernames,fp)\n",
    "    \n",
    "# with open('./Dataset/user_info.pkl','rb') as fp:\n",
    "#         user_tweet_info = pickle.load(fp)\n",
    "\n",
    "# user_tweet_info_hate = {}\n",
    "# user_tweet_info_counter = {}\n",
    "# for name,tweet_info in user_tweet_info.items():\n",
    "#     if name in hate :\n",
    "#         user_tweet_info_hate.update({name:tweet_info})\n",
    "#     elif name in counter:\n",
    "#         user_tweet_info_counter.update({name:tweet_info})\n",
    "\n",
    "# with open('user_tweet_info_counter.pkl','wb') as fp :\n",
    "#     pickle.dump(user_tweet_info_counter,fp)\n",
    "# with open('user_tweet_info_hate.pkl','wb') as fp :\n",
    "#     pickle.dump(user_tweet_info_hate,fp)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hashtag Based Analysis :"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* __hashtag_words_Hate__ : Dictionary with keys as __hate users__ names and values equal to list of all hashtag words posted by that particlar user. \n",
    "* __hashtag_words_Counter__ :Dictionary with keys as __Counter users__ names and values equal to list of all hashtag words posted by that particlar user."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:43:40.589572Z",
     "start_time": "2018-10-29T01:43:38.235595Z"
    }
   },
   "outputs": [],
   "source": [
    "hashtag_words_Hate = []\n",
    "hashtag_words_Counter = []    \n",
    "hashtag_words_NonCounter = []\n",
    "\n",
    "######################################################\n",
    "\n",
    "hashtag_words_Hate_userwise = {}\n",
    "hashtag_words_Counter_userwise = {}\n",
    "hashtag_words_NonCounter_userwise = {}\n",
    "######################################################\n",
    "\n",
    "hashtags_per_tweet_hate = []\n",
    "hashtags_per_tweet_counter = []\n",
    "hashtags_per_tweet_NonCounter = []\n",
    "######################################################\n",
    "\n",
    "# hate_users_tweets_less_than_10 = []\n",
    "# counter_users_tweets_less_than_10 = []\n",
    "\n",
    "for name,tweets in all_user_tweets.items():\n",
    "    ind_tweets = ' '.join(tweets)    ##### all tweets for particular user\n",
    "    if name.strip() in hate_usernames:\n",
    "        hashtag_words_Hate.append(' '.join([i for i in set(re.findall(r'\\w+',' '.join(re.findall(r'#\\w+',ind_tweets))))]))\n",
    "        hashtag_words_Hate_userwise[name] = re.findall(r'\\w+',' '.join(re.findall(r'#\\w+',ind_tweets)))\n",
    "        hashtags_per_tweet_hate.append(len(hashtag_words_Hate_userwise[name])/len(tweets))\n",
    "#         if len(tweets)<10:\n",
    "#             hate_users_tweets_less_than_10.append(name)\n",
    "            \n",
    "    elif name.strip() in counter_usernames:\n",
    "        hashtag_words_Counter.append(' '.join([i for i in set(re.findall(r'\\w+',' '.join(re.findall(r'#\\w+',ind_tweets))))]))\n",
    "        hashtag_words_Counter_userwise[name] = re.findall(r'\\w+',' '.join(re.findall(r'#\\w+',ind_tweets))) \n",
    "        hashtags_per_tweet_counter.append(len(hashtag_words_Counter_userwise[name])/len(tweets))\n",
    "#         if len(tweets)<10:\n",
    "#             counter_users_tweets_less_than_10.append(name)\n",
    "    elif name.strip() in Non_counter_usernames:\n",
    "        hashtag_words_NonCounter.append(' '.join([i for i in set(re.findall(r'\\w+',' '.join(re.findall(r'#\\w+',ind_tweets))))]))\n",
    "        hashtag_words_NonCounter_userwise[name] = re.findall(r'\\w+',' '.join(re.findall(r'#\\w+',ind_tweets))) \n",
    "        hashtags_per_tweet_NonCounter.append(len(hashtag_words_NonCounter_userwise[name])/len(tweets))\n",
    "hashtag_words_Hate_wordcloud = np.array(sorted(Counter(re.findall(r'\\w+',(' '.join(hashtag_words_Hate)).lower())).items(),key = lambda x : x[1],reverse = True))[:200,0]\n",
    "hashtag_words_Counter_wordcloud = np.array(sorted(Counter(re.findall(r'\\w+',(' '.join(hashtag_words_Counter)).lower())).items(),key = lambda x : x[1],reverse = True))[:200,0]\n",
    "hashtag_words_NonCounter_wordcloud = np.array(sorted(Counter(re.findall(r'\\w+',(' '.join(hashtag_words_NonCounter)).lower())).items(),key = lambda x : x[1],reverse = True))[:200,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:43:40.594045Z",
     "start_time": "2018-10-29T01:43:40.591468Z"
    }
   },
   "outputs": [],
   "source": [
    "# with open('Refined_Project_Dataset/hashtag_words_Hate_wordcloud.json','w') as fp:\n",
    "#     json.dump(sorted(Counter(re.findall(r'[a-zA-Z]+',' '.join(hashtag_words_Hate))).items(),key = lambda x : x[1],reverse = True),fp)\n",
    "# with open('Refined_Project_Dataset/hashtag_words_Counter_wordcloud.json','w') as fp:\n",
    "#     json.dump(sorted(Counter(re.findall(r'[a-zA-Z]+',' '.join(hashtag_words_Counter))).items(),key = lambda x : x[1],reverse = True),fp)\n",
    "# with open('Refined_Project_Dataset/hashtag_words_NonCounter_wordcloud.json','w') as fp:\n",
    "#     json.dump(sorted(Counter(re.findall(r'[a-zA-Z]+',' '.join(hashtag_words_NonCounter))).items(),key = lambda x : x[1],reverse = True),fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:36:12.719657Z",
     "start_time": "2018-10-23T21:36:12.711900Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(200, 200, 200)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# np.savetxt('hashtag_words_Hate_wordcloud.txt',[hashtag_words_Hate_wordcloud],fmt = '%s')\n",
    "# np.savetxt('hashtag_words_Counter_wordcloud.txt',[hashtag_words_Counter_wordcloud],fmt = '%s')\n",
    "# np.savetxt('hashtag_words_NonCounter_wordcloud.txt',[hashtag_words_NonCounter_wordcloud],fmt = '%s')\n",
    "len(hashtag_words_Hate_wordcloud), len(hashtag_words_Counter_wordcloud),len(hashtag_words_NonCounter_wordcloud)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:36:14.077878Z",
     "start_time": "2018-10-23T21:36:14.070999Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['1', 'newprofilepic', '2', 'worldcup', 'tbt', 'breaking', 'metoo',\n",
       "       'blackpanther', 'trump', 'netneutrality'], dtype='<U58')"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hashtag_words_NonCounter_wordcloud[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* __hashtag_words_Hate_wordcloud__ : List of all hashtag words posted by all __Hate users__. \n",
    "* __hashtag_words_Counter_wordcloud__ :List of all hashtag words posted by all __Counter users__.\n",
    "* __hashtag_words_NonCounter_wordcloud__ :List of all hashtag words posted by all __NonCounter users__."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### T-Test/Mann-Whitney U-Test and Boxplot :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T22:24:01.425703Z",
     "start_time": "2018-10-23T22:24:00.918887Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hate Vs Counter\n",
      "Ttest_indResult(statistic=4.13489708078942, pvalue=3.7283125451743634e-05)\n",
      "MannwhitneyuResult(statistic=243545.5, pvalue=1.1986761645065478e-08)\n",
      "\n",
      "Hate Vs NonCounter\n",
      "Ttest_indResult(statistic=1.4694908541389018, pvalue=0.1421522072031254)\n",
      "MannwhitneyuResult(statistic=40673.5, pvalue=0.001183110248419422)\n",
      "\n",
      "Counter Vs NonCounter\n",
      "Ttest_indResult(statistic=1.4197946271126023, pvalue=0.15589990141702753)\n",
      "MannwhitneyuResult(statistic=104492.0, pvalue=0.20416659296044926)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 158.4x144 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 158.4x144 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "['#4c72b0', '#55a868', '#c44e52', '#8172b2', '#ccb974', '#64b5cd']"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sns.set()\n",
    "Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(hashtags_per_tweet_NonCounter))],hashtags_per_tweet_NonCounter],columns = ['Users',''])\n",
    "counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(hashtags_per_tweet_counter))],hashtags_per_tweet_counter],columns = ['Users',''])\n",
    "hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(hashtags_per_tweet_hate))],hashtags_per_tweet_hate],columns = ['Users',''])\n",
    "Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "print('Hate Vs Counter')\n",
    "print(ttest_ind(hashtags_per_tweet_counter,hashtags_per_tweet_hate))\n",
    "print(mannwhitneyu(hashtags_per_tweet_counter,hashtags_per_tweet_hate),end = '\\n\\n')\n",
    "print('Hate Vs NonCounter')\n",
    "print(ttest_ind(hashtags_per_tweet_NonCounter,hashtags_per_tweet_hate))\n",
    "print(mannwhitneyu(hashtags_per_tweet_NonCounter,hashtags_per_tweet_hate),end = '\\n\\n')\n",
    "print('Counter Vs NonCounter')\n",
    "print(ttest_ind(hashtags_per_tweet_counter,hashtags_per_tweet_NonCounter))\n",
    "print(mannwhitneyu(hashtags_per_tweet_counter,hashtags_per_tweet_NonCounter))\n",
    "# palette = {'Counter':'darkslateblue','Hate':'crimson','NonCounter':'forestgreen'}\n",
    "plt.figure(figsize=(2.2,2))\n",
    "ax = sns.boxplot(x = 'Users',y ='',data = Comb_df ,showfliers = False, linewidth=1,width=0.5,palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "ax.set_xticklabels('')\n",
    "plt.xlabel('')\n",
    "plt.title('hashtags/tweet')\n",
    "plt.tight_layout()\n",
    "plt.savefig('Images/hashtagsPERtweet.eps', format='eps')\n",
    "plt.show()\n",
    "############################### Barplot#############################\n",
    "plt.figure(figsize=(2.2,2))\n",
    "ax = sns.barplot(x = 'Users',y ='',data = Comb_df,estimator = np.median,errwidth=1.2 )\n",
    "ax.set_xticklabels('')\n",
    "plt.title('hashtags/tweet')\n",
    "plt.xlabel('')\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "current_palette = sns.color_palette()\n",
    "current_palette.as_hex()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Lexical and Sentiment analysis on hatespeakes and CounterSpeaker Tweets: "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Below __Code snippet__ implements __lexicon_counter_user__,__lexicon_hate_user__,__sentiment_counter_user__ ,__sentiment_hate_user__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* __Lexical Analysis and TextBlob:__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Interest Topics :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-03T01:48:45.921154Z",
     "start_time": "2018-11-03T01:48:45.917742Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(548, 1239)"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(hate_usernames),len(counter_usernames)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-03T02:37:41.692843Z",
     "start_time": "2018-11-03T02:37:40.713358Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 3/3 [00:00<00:00, 26.16it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "all--1637\n",
      "popular--823\n",
      "niche--4573\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "100%|██████████| 3/3 [00:00<00:00,  8.95it/s]\n"
     ]
    }
   ],
   "source": [
    "from tqdm import tqdm\n",
    "with open('output328.json','r') as fp:\n",
    "    output = json.load(fp)\n",
    "len_hate = len(hate_usernames.intersection(set(output.keys()))) \n",
    "len_counter = len((counter_usernames.union(Non_counter_usernames)).intersection(set(output.keys()))) \n",
    "\n",
    "hateuser_interests = {}\n",
    "Counteruser_interests = {}\n",
    "for user in output.keys():\n",
    "    try:\n",
    "        if user in hate_usernames:\n",
    "            hateuser_interests[user]={}\n",
    "            for each_category_list in output[user]['response']:\n",
    "                hateuser_interests[user][each_category_list[0]] = {}\n",
    "                sum_  = sum(list(each_category_list[1].values()))\n",
    "                for key,value in each_category_list[1].items():\n",
    "                    hateuser_interests[user][each_category_list[0]].update({key : value/sum_})\n",
    "        if user in counter_usernames:\n",
    "            Counteruser_interests[user]={}\n",
    "            for each_category_list in output[user]['response']:\n",
    "                Counteruser_interests[user][each_category_list[0]] = {}\n",
    "                sum_  = sum(list(each_category_list[1].values()))\n",
    "                for key,value in each_category_list[1].items():\n",
    "                    Counteruser_interests[user][each_category_list[0]].update({key : value/sum_})\n",
    "    except:\n",
    "        pass\n",
    "#------------------------------------------------------------------------------------------\n",
    "hate = {}\n",
    "for dict_ in hateuser_interests.values():\n",
    "    for class_,values_dict in dict_.items(): \n",
    "        if class_ not in hate.keys():\n",
    "                hate[class_] = {}\n",
    "        else:\n",
    "            for name,value in values_dict.items():\n",
    "                if name not in hate[class_].keys():\n",
    "                    hate[class_][name] = []\n",
    "                else:\n",
    "                    hate[class_][name].append(value)\n",
    "counter = {}\n",
    "for dict_ in Counteruser_interests.values():\n",
    "    for class_,values_dict in dict_.items(): \n",
    "        if class_ not in counter.keys():\n",
    "                counter[class_] = {}\n",
    "        else:\n",
    "            for name,value in values_dict.items():\n",
    "                if name not in counter[class_].keys():\n",
    "                    counter[class_][name] = []\n",
    "                else:\n",
    "                    counter[class_][name].append(value)\n",
    "#--------------------------------------------------------------------\n",
    "vocab = {}\n",
    "for class_ in hate.keys():\n",
    "    vocab[class_]=[i for i in set(list(hate[class_].keys())+list(counter[class_].keys()))]\n",
    "    print('{}--{}'.format(class_,len(vocab[class_])))\n",
    "#-------------------------------------------------------\n",
    "# Binny asked Dictionary:\n",
    "#Hate\n",
    "for class_,topic_dict in tqdm(hate.items()):\n",
    "    for topic,value in topic_dict.items():\n",
    "                if hate[class_][topic]==[]:\n",
    "                    hate[class_][topic].extend(np.zeros(len(hate_usernames)).tolist())\n",
    "                else:\n",
    "                    hate[class_][topic].extend(np.zeros(len(hate_usernames)-len(hate[class_][topic])).tolist())\n",
    "for class_,topics in vocab.items():\n",
    "    for name in topics:\n",
    "        if name not in hate[class_].keys():\n",
    "            hate[class_][name] = np.zeros(len(hate_usernames)).tolist()\n",
    "#Counter            \n",
    "for class_,topic_dict in tqdm(counter.items()):\n",
    "    for topic,value in topic_dict.items():\n",
    "                if counter[class_][topic]==[]:\n",
    "                    counter[class_][topic].extend(np.zeros(len(counter_usernames)).tolist())\n",
    "                else:\n",
    "                    counter[class_][topic].extend(np.zeros(len(counter_usernames)-len(counter[class_][topic])).tolist())\n",
    "for class_,topics in vocab.items():\n",
    "    for name in topics:\n",
    "        if name not in counter[class_].keys():\n",
    "            counter[class_][name] = np.zeros(len(counter_usernames)).tolist()\n",
    "# with open('Refined_Project_Dataset/Binny_Hate_interest_dict.json','w') as fp:\n",
    "#     json.dump(hate,fp)\n",
    "# with open('Refined_Project_Dataset/Binny_Counter_interest_dict.json','w') as fp:\n",
    "#     json.dump(counter,fp)\n",
    "# with open('Refined_Project_Dataset/Binny_Hate_interest_dict.json','r') as fp:\n",
    "#     hate = json.load(fp)\n",
    "# with open('Refined_Project_Dataset/Binny_Counter_interest_dict.json','r') as fp:\n",
    "#     counter = json.load(fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-03T05:06:08.494693Z",
     "start_time": "2018-11-03T05:06:05.334822Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**************************** all *******************************\n",
      "bbc\n",
      "Hate-Mean: 0.00015659421075130714--Median: 0.0 | Counter-Mean: 1.7764299255374202e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337283.0, pvalue=0.008345760341552258)\n",
      "\n",
      "ncaa\n",
      "Hate-Mean: 0.00015719560992236592--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337627.5, pvalue=0.004591739919921564)\n",
      "\n",
      "nba players\n",
      "Hate-Mean: 0.0004467558107353772--Median: 0.0 | Counter-Mean: 0.0001380229936761032--Median: 0.0\n",
      "MannwhitneyuResult(statistic=335564.5, pvalue=0.006813612204461781)\n",
      "\n",
      "niggas\n",
      "Hate-Mean: 0.0004345352082767713--Median: 0.0 | Counter-Mean: 0.00010200964160115918--Median: 0.0\n",
      "MannwhitneyuResult(statistic=335016.0, pvalue=0.0012683928275952887)\n",
      "\n",
      "shit\n",
      "Hate-Mean: 0.0014694689615932922--Median: 0.0 | Counter-Mean: 0.0008069067061101916--Median: 0.0\n",
      "MannwhitneyuResult(statistic=331599.5, pvalue=0.008206764033900681)\n",
      "\n",
      "tube\n",
      "Hate-Mean: 0.0011194900705917643--Median: 0.0 | Counter-Mean: 0.00039864411667646924--Median: 0.0\n",
      "MannwhitneyuResult(statistic=333055.0, pvalue=0.0026908567572964492)\n",
      "\n",
      "hoops\n",
      "Hate-Mean: 0.0003940709935424973--Median: 0.0 | Counter-Mean: 0.00011986803983292157--Median: 0.0\n",
      "MannwhitneyuResult(statistic=335556.5, pvalue=0.006718431021018907)\n",
      "\n",
      "country music\n",
      "Hate-Mean: 0.0002596460183859339--Median: 0.0 | Counter-Mean: 2.5487447432139673e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337281.5, pvalue=0.00830874627694689)\n",
      "\n",
      "professional\n",
      "Hate-Mean: 0.00032502025377773887--Median: 0.0 | Counter-Mean: 7.257904011838336e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336240.5, pvalue=0.008622641378523204)\n",
      "\n",
      "timeline\n",
      "Hate-Mean: 0.000728045630302709--Median: 0.0 | Counter-Mean: 0.00010513859534528947--Median: 0.0\n",
      "MannwhitneyuResult(statistic=333409.5, pvalue=0.0001635103560464963)\n",
      "\n",
      "funny shit\n",
      "Hate-Mean: 0.0001533067661863423--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337627.5, pvalue=0.004591739919921564)\n",
      "\n",
      "monster brasil\n",
      "Hate-Mean: 0.00013129636303512132--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337627.5, pvalue=0.004591739919921564)\n",
      "\n",
      "**************************** popular *******************************\n",
      "texas\n",
      "Hate-Mean: 0.00035640081996847565--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337008.0, pvalue=0.0013096415212768273)\n",
      "\n",
      "entertainers\n",
      "Hate-Mean: 0.0002459841849855671--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337627.5, pvalue=0.004591739919921564)\n",
      "\n",
      "shit\n",
      "Hate-Mean: 0.002018665357662875--Median: 0.0 | Counter-Mean: 0.0010598515419282524--Median: 0.0\n",
      "MannwhitneyuResult(statistic=328595.5, pvalue=0.0019404272673881205)\n",
      "\n",
      "timeline\n",
      "Hate-Mean: 0.0006396051292161534--Median: 0.0 | Counter-Mean: 5.6335422280255124e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=334103.0, pvalue=0.00013866532475585058)\n",
      "\n",
      "basket\n",
      "Hate-Mean: 0.00023534539183515734--Median: 0.0 | Counter-Mean: 6.332605381322178e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336597.0, pvalue=0.009584633525915515)\n",
      "\n",
      "cute\n",
      "Hate-Mean: 0.0002948284553678427--Median: 0.0 | Counter-Mean: 2.7798809086524883e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336931.5, pvalue=0.009463659472743741)\n",
      "\n",
      "**************************** niche *******************************\n",
      "miscellaneous\n",
      "Hate-Mean: 0.0013833684742526942--Median: 0.0 | Counter-Mean: 0.0004609332617383561--Median: 0.0\n",
      "MannwhitneyuResult(statistic=329181.0, pvalue=6.77115834103104e-05)\n",
      "\n",
      "crazy\n",
      "Hate-Mean: 0.00042866530325805864--Median: 0.0 | Counter-Mean: 5.339497142927747e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=335696.5, pvalue=0.00106165730088776)\n",
      "\n",
      "homies\n",
      "Hate-Mean: 0.0009128807238996187--Median: 0.0 | Counter-Mean: 0.00014266253214530978--Median: 0.0\n",
      "MannwhitneyuResult(statistic=332094.0, pvalue=5.965606269087396e-05)\n",
      "\n",
      "mother\n",
      "Hate-Mean: 0.00027523067962832037--Median: 0.0 | Counter-Mean: 5.8361909316980606e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336936.5, pvalue=0.009580987707312423)\n",
      "\n",
      "adult industry\n",
      "Hate-Mean: 0.0004252702561764007--Median: 0.0 | Counter-Mean: 6.664826426759589e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336317.0, pvalue=0.0032213360951714198)\n",
      "\n",
      "guns\n",
      "Hate-Mean: 0.0002153769728072398--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337627.5, pvalue=0.004591739919921564)\n",
      "\n",
      "sneakers\n",
      "Hate-Mean: 0.0009800428587325425--Median: 0.0 | Counter-Mean: 0.00029110661586930544--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336243.0, pvalue=0.008665685449941514)\n",
      "\n",
      "nicki\n",
      "Hate-Mean: 0.0005129008742357085--Median: 0.0 | Counter-Mean: 0.00011473515960056339--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336939.5, pvalue=0.009651997652127821)\n",
      "\n",
      "niggas\n",
      "Hate-Mean: 0.005556173606196713--Median: 0.0 | Counter-Mean: 0.0031828283172332748--Median: 0.0\n",
      "MannwhitneyuResult(statistic=328055.5, pvalue=0.008369154483469044)\n",
      "\n",
      "tgod\n",
      "Hate-Mean: 0.0002965334614300448--Median: 0.0 | Counter-Mean: 6.120059347473878e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336589.5, pvalue=0.009429546583194958)\n",
      "\n",
      "ufc fighters\n",
      "Hate-Mean: 0.0006547303143139519--Median: 0.0 | Counter-Mean: 0.00025714908837085396--Median: 0.0\n",
      "MannwhitneyuResult(statistic=335979.0, pvalue=0.003485873074390828)\n",
      "\n",
      "official nba\n",
      "Hate-Mean: 0.00015585494545216857--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337008.0, pvalue=0.0013096415212768273)\n",
      "\n",
      "film news\n",
      "Hate-Mean: 0.0003058165009460284--Median: 0.0 | Counter-Mean: 1.8343238682221733e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337280.0, pvalue=0.008271883239012729)\n",
      "\n",
      "travel companies\n",
      "Hate-Mean: 0.00021970109973406418--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337008.0, pvalue=0.0013096415212768273)\n",
      "\n",
      "basketball association\n",
      "Hate-Mean: 0.00017836954528810604--Median: 0.0 | Counter-Mean: 2.9892685259916896e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337284.0, pvalue=0.008370514458090327)\n",
      "\n",
      "phone apps\n",
      "Hate-Mean: 0.000237420105703954--Median: 0.0 | Counter-Mean: 4.0355125100887815e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337284.0, pvalue=0.008370514458090327)\n",
      "\n",
      "sports teams\n",
      "Hate-Mean: 0.0010235331141664685--Median: 0.0 | Counter-Mean: 0.000337254829252043--Median: 0.0\n",
      "MannwhitneyuResult(statistic=333821.5, pvalue=0.0028192646761579553)\n",
      "\n",
      "fifth\n",
      "Hate-Mean: 0.00039600666464283585--Median: 0.0 | Counter-Mean: 5.660108557348651e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336585.0, pvalue=0.00933755826133817)\n",
      "\n",
      "sprint cup\n",
      "Hate-Mean: 0.0001595415549026253--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337627.5, pvalue=0.004591739919921564)\n",
      "\n",
      "nbateams\n",
      "Hate-Mean: 0.00026506248534005665--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337008.0, pvalue=0.0013096415212768273)\n",
      "\n",
      "nba news\n",
      "Hate-Mean: 0.0020427683958896482--Median: 0.0 | Counter-Mean: 0.0007337389533227586--Median: 0.0\n",
      "MannwhitneyuResult(statistic=334266.0, pvalue=0.009853465348204277)\n",
      "\n",
      "hoop\n",
      "Hate-Mean: 0.00015537700955144397--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337008.0, pvalue=0.0013096415212768273)\n",
      "\n",
      "report\n",
      "Hate-Mean: 0.0003222994795219152--Median: 0.0 | Counter-Mean: 5.29527567461385e-05--Median: 0.0\n",
      "MannwhitneyuResult(statistic=336319.5, pvalue=0.0032423650159838224)\n",
      "\n",
      "rihanna\n",
      "Hate-Mean: 0.001121244391483717--Median: 0.0 | Counter-Mean: 0.00020153687787392838--Median: 0.0\n",
      "MannwhitneyuResult(statistic=333891.0, pvalue=0.0017948448264161277)\n",
      "\n",
      "ceos\n",
      "Hate-Mean: 0.00015168849006814094--Median: 0.0 | Counter-Mean: 0.0--Median: 0.0\n",
      "MannwhitneyuResult(statistic=337627.5, pvalue=0.004591739919921564)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for interest,classes in vocab.items():\n",
    "    print('****************************',interest,'*******************************')\n",
    "    for key in classes:\n",
    "#         print(counter[interest][key])\n",
    "        try:\n",
    "            _,pvalue = mannwhitneyu(hate[interest][key],counter[interest][key])\n",
    "            if pvalue<=0.01 and np.mean(hate[interest][key])>np.mean(counter[interest][key]) :\n",
    "                print(key)\n",
    "                print('Hate-Mean: {}--Median: {} | Counter-Mean: {}--Median: {}'.format(np.mean(hate[interest][key]),np.median(hate[interest][key]),np.mean(counter[interest][key]),np.median(counter[interest][key])))\n",
    "                print(mannwhitneyu(hate[interest][key],counter[interest][key]),end = '\\n\\n')\n",
    "        except:\n",
    "            pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:53:33.394256Z",
     "start_time": "2018-10-23T21:53:31.428013Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# lexicon_hate_user = []\n",
    "# lexicon_counter_user = []\n",
    "# # lexicon_Noncounter_user = []\n",
    "# for name,tweets in all_user_tweets.items():\n",
    "#     ind_tweets = ' '.join(tweets)    ###### all tweets for particular user\n",
    "#     if name.strip() in hate_usernames:\n",
    "#         lexicon_hate_user.append([ind_tweets,1])\n",
    "# #         lexicon_hate_user[name] = lexicon.analyze(ind_tweets,normalize = True)\n",
    "#     elif name.strip() in list(counter_usernames)+list(Non_counter_usernames):\n",
    "#         lexicon_counter_user.append([ind_tweets,0])\n",
    "#         lexicon_counter_user[name] = lexicon.analyze(ind_tweets,normalize = True)\n",
    "#     elif name.strip() in Non_counter_usernames:\n",
    "#         lexicon_Noncounter_user[name] = lexicon.analyze(ind_tweets,normalize = True)\n",
    "#-----------------------------------------------------------------------------------------------------------------#\n",
    "# from textblob import TextBlob\n",
    "# textblob_hate_user = {}\n",
    "# textblob_counter_user = {}\n",
    "# textblob_Noncounter_user = {}\n",
    "# for counter,(name,tweets) in enumerate(all_user_tweets.items()):\n",
    "#     print(counter)\n",
    "#     ind_tweets = ' '.join(tweets)    ###### all tweets for particular user\n",
    "#     if name.strip() in hate_usernames:\n",
    "#         textblob_hate_user[name] = {'polarity':TextBlob(ind_tweets).sentiment.polarity,'subjectivity':TextBlob(ind_tweets).sentiment.subjectivity}\n",
    "#     elif name.strip() in counter_usernames:\n",
    "#         textblob_counter_user[name] = {'polarity':TextBlob(ind_tweets).sentiment.polarity,'subjectivity':TextBlob(ind_tweets).sentiment.subjectivity}\n",
    "#     elif name.strip() in Non_counter_usernames:\n",
    "#         textblob_Noncounter_user[name] = {'polarity':TextBlob(ind_tweets).sentiment.polarity,'subjectivity':TextBlob(ind_tweets).sentiment.subjectivity}\n",
    "# --------------------------------------------------------------------------------------------------------\n",
    "# with open('All_hate_tweets_with_labels.pkl','wb') as fp:\n",
    "#     pickle.dump(lexicon_hate_user,fp)\n",
    "# with open('All_counter_tweets_with_labels.pkl','wb') as fp:\n",
    "#     pickle.dump(lexicon_counter_user,fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:36:32.980858Z",
     "start_time": "2018-10-23T21:36:32.975010Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "509"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(textblob_hate_user)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* __Sentiment Analysis __/__Profane Words Distribution__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:36:33.479154Z",
     "start_time": "2018-10-23T21:36:33.469644Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Profane_Words = np.loadtxt('Refined_Project_Dataset/ProfaneWords.csv',dtype = str,delimiter = ',')\n",
    "# profane_hate_user = []\n",
    "# profane_counter_user = []\n",
    "# profane_Noncounter_user = []\n",
    "# # profane_user_dict = {}\n",
    "# # sentiment_hate_user = {}\n",
    "# # sentiment_counter_user = {}\n",
    "# # sentiment_Noncounter_user = {}\n",
    "# for line_counter,(name,tweets) in enumerate(all_user_tweets.items()):\n",
    "# #     ind_tweets = ' '.join(tweets)\n",
    "#     counter = 0\n",
    "#     if name.strip() in hate_usernames:\n",
    "#         print(line_counter,name)\n",
    "# #         Hneg,Hneu,Hpos,Hcomp = 0,0,0,0\n",
    "#         for tweet in tweets:\n",
    "#             for word in Profane_Words:\n",
    "#                 if len(re.findall(word,tweet.lower()))!=0:\n",
    "#                     counter+=1\n",
    "#                     break\n",
    "#         print(counter)\n",
    "#         profane_hate_user.append(counter/len(tweets))\n",
    "# #         print(profane_user_dict)\n",
    "# #             sentiment = sentiment_analyzer.polarity_scores(each_tweet)\n",
    "# #             Hneg += sentiment['neg']\n",
    "# #             Hneu += sentiment['neu']\n",
    "# #             Hpos += sentiment['pos']\n",
    "# #             Hcomp += sentiment['compound']\n",
    "# #         sentiment_hate_user[name] = {'neg':Hneg/len(tweets),'neu': Hneu/len(tweets),'pos':Hpos/len(tweets),'compound':Hcomp/len(tweets)}\n",
    "        \n",
    "#     elif name.strip() in counter_usernames:\n",
    "#         print(line_counter,name)\n",
    "#         for tweet in tweets:\n",
    "#             for word in Profane_Words:\n",
    "#                 if len(re.findall(word,tweet.lower()))!=0:\n",
    "#                     counter+=1\n",
    "#                     break\n",
    "#         print(counter)\n",
    "#         profane_counter_user.append(counter/len(tweets))\n",
    "# #         Cneg,Cneu,Cpos,Ccomp = 0,0,0,0\n",
    "# #         for word in Profane_Words:\n",
    "# #             profane_user_dict[word] = float(len(re.findall(word,ind_tweets.lower())))/len(tweets)\n",
    "# # #             print(len(re.findall(word,ind_tweets.lower())),end ='.')\n",
    "# #         profane_counter_user[name]=  profane_user_dict\n",
    "# #             sentiment = sentiment_analyzer.polarity_scores(each_tweet)\n",
    "# #             Cneg += sentiment['neg']\n",
    "# #             Cneu += sentiment['neu']\n",
    "# #             Cpos += sentiment['pos']\n",
    "# #             Ccomp += sentiment['compound']\n",
    "# #         sentiment_counter_user[name] = {'neg':Cneg/len(tweets),'neu': Cneu/len(tweets),'pos':Cpos/len(tweets),'compound':Ccomp/len(tweets)}\n",
    "#     elif name.strip() in Non_counter_usernames:\n",
    "#         print(line_counter,name)\n",
    "#         for tweet in tweets:\n",
    "#             for word in Profane_Words:\n",
    "#                 if len(re.findall(word,tweet.lower()))!=0:\n",
    "#                     counter+=1\n",
    "#                     break\n",
    "#         print(counter)\n",
    "#         profane_Noncounter_user.append(counter/len(tweets))\n",
    "# #         NCneg,NCneu,NCpos,NCcomp = 0,0,0,0\n",
    "# #         for word in Profane_Words:\n",
    "# #             profane_user_dict[word] = float(len(re.findall(word,ind_tweets.lower())))/len(tweets)\n",
    "# # #             print(len(re.findall(word,ind_tweets.lower())),end ='.')\n",
    "# #         profane_Noncounter_user[name]=  profane_user_dict\n",
    "# #             sentiment = sentiment_analyzer.polarity_scores(each_tweet)\n",
    "# #             NCneg += sentiment['neg']\n",
    "# #             NCneu += sentiment['neu']\n",
    "# #             NCpos += sentiment['pos']\n",
    "# #             NCcomp += sentiment['compound']\n",
    "# #         sentiment_Noncounter_user[name] = {'neg':NCneg/len(tweets),'neu': NCneu/len(tweets),'pos':NCpos/len(tweets),'compound':NCcomp/len(tweets)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-01T23:12:55.991491Z",
     "start_time": "2018-11-01T23:12:55.976740Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.15257017355756078, 0.13237973840471354, 0.1603729497418059)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "####### Profane mean\n",
    "np.mean(profane_hate_user),np.mean(profane_counter_user),np.mean(profane_Noncounter_user)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-01T23:12:56.248800Z",
     "start_time": "2018-11-01T23:12:56.242435Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.14411302982731555, 0.11881031792057731, 0.149438202247191)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "####### Profane median\n",
    "np.median(profane_hate_user),np.median(profane_counter_user),np.median(profane_Noncounter_user)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T22:27:17.291222Z",
     "start_time": "2018-10-23T22:27:17.125164Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hate----Counter\n",
      "MannwhitneyuResult(statistic=1154.0, pvalue=0.047741239297343474)\n",
      "\n",
      "Counter----NonCounter\n",
      "MannwhitneyuResult(statistic=609.0, pvalue=0.031434494068282796)\n",
      "\n",
      "Hate----NonCounter\n",
      "MannwhitneyuResult(statistic=234.0, pvalue=0.3923935434248835)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 158.4x144 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print('Hate----Counter')\n",
    "print(mannwhitneyu(profane_hate_user,profane_counter_user),end='\\n\\n')\n",
    "print('Counter----NonCounter')\n",
    "print(mannwhitneyu(profane_Noncounter_user,profane_counter_user),end='\\n\\n')\n",
    "print('Hate----NonCounter')\n",
    "print(mannwhitneyu(profane_Noncounter_user,profane_hate_user))\n",
    "######################## Profane Images ############################\n",
    "Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(profane_Noncounter_user))],profane_Noncounter_user],columns = ['Users',''])\n",
    "counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(profane_counter_user))],profane_counter_user],columns = ['Users',''])\n",
    "hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(profane_hate_user))],profane_hate_user],columns = ['Users',''])\n",
    "Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "plt.figure(figsize = (2.2,2))\n",
    "ax  = sns.boxplot(x = 'Users',y ='',data = Comb_df, linewidth=1,width=0.6,showfliers = False,palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "ax.set_xticklabels('')\n",
    "ax.set_xlabel('')\n",
    "ax.set_title('Profane')\n",
    "plt.tight_layout()\n",
    "plt.savefig('Images/Profane.eps', format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### This code is for writing the above created lexicon and sentiment dictionaries into pkl files so as speed up their importing as well as saving memory : "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:36:44.102215Z",
     "start_time": "2018-10-23T21:36:44.096514Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'polarity': 0.11299070730567451, 'subjectivity': 0.5946555603785325}"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(textblob_hate_user.values())[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-04T01:59:07.589420Z",
     "start_time": "2018-10-04T01:59:07.582962Z"
    }
   },
   "outputs": [],
   "source": [
    "# with open('Refined_Project_Dataset/lexicon_hate_user.pkl','wb') as fp:\n",
    "#     pickle.dump(lexicon_hate_user,fp)\n",
    "# with open('Refined_Project_Dataset/lexicon_counter_user.pkl','wb') as fp:\n",
    "#     pickle.dump(lexicon_counter_user,fp)\n",
    "# with open('Refined_Project_Dataset/lexicon_Noncounter_user.pkl','wb') as fp:\n",
    "#     pickle.dump(lexicon_Noncounter_user,fp)\n",
    "# with open('Refined_Project_Dataset/textblob_hate_user.pkl','wb') as fp:\n",
    "#     pickle.dump(textblob_hate_user,fp)\n",
    "# with open('Refined_Project_Dataset/textblob_counter_user.pkl','wb') as fp:\n",
    "#     pickle.dump(textblob_counter_user,fp)\n",
    "# with open('Refined_Project_Dataset/textblob_Noncounter_user.pkl','wb') as fp:\n",
    "#     pickle.dump(textblob_Noncounter_user,fp)\n",
    "# with open('Refined_Project_Dataset/sentiment_hate_user.pkl','wb') as fp:\n",
    "#     pickle.dump(sentiment_hate_user,fp)\n",
    "# with open('Refined_Project_Dataset/sentiment_counter_user.pkl','wb') as fp:\n",
    "#     pickle.dump(sentiment_counter_user,fp)\n",
    "# with open('Refined_Project_Dataset/sentiment_Noncounter_user.pkl','wb') as fp:\n",
    "# #     pickle.dump(sentiment_Noncounter_user,fp)\n",
    "# with open('Refined_Project_Dataset/profane_hate_user.pkl','wb') as fp:\n",
    "#     pickle.dump(profane_hate_user,fp)\n",
    "# with open('Refined_Project_Dataset/profane_counter_user.pkl','wb') as fp:\n",
    "#     pickle.dump(profane_counter_user,fp)\n",
    "# with open('Refined_Project_Dataset/profane_Noncounter_user.pkl','wb') as fp:\n",
    "#     pickle.dump(profane_Noncounter_user,fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:36:47.395119Z",
     "start_time": "2018-10-23T21:36:47.390892Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(188, 1155)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(textblob_Noncounter_user.keys()), len(textblob_counter_user.keys())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating a lexicon DataFrame for all Users : "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:43:52.294688Z",
     "start_time": "2018-10-29T01:43:52.136332Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1852, 194)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>achievement</th>\n",
       "      <th>affection</th>\n",
       "      <th>aggression</th>\n",
       "      <th>air_travel</th>\n",
       "      <th>alcohol</th>\n",
       "      <th>ancient</th>\n",
       "      <th>anger</th>\n",
       "      <th>animal</th>\n",
       "      <th>anonymity</th>\n",
       "      <th>anticipation</th>\n",
       "      <th>...</th>\n",
       "      <th>wealthy</th>\n",
       "      <th>weapon</th>\n",
       "      <th>weather</th>\n",
       "      <th>wedding</th>\n",
       "      <th>white_collar_job</th>\n",
       "      <th>work</th>\n",
       "      <th>worship</th>\n",
       "      <th>writing</th>\n",
       "      <th>youth</th>\n",
       "      <th>zest</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>crasykayla</th>\n",
       "      <td>0.001277</td>\n",
       "      <td>0.003922</td>\n",
       "      <td>0.000578</td>\n",
       "      <td>0.000304</td>\n",
       "      <td>0.002675</td>\n",
       "      <td>0.000426</td>\n",
       "      <td>0.000334</td>\n",
       "      <td>0.001186</td>\n",
       "      <td>0.000061</td>\n",
       "      <td>0.000061</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000638</td>\n",
       "      <td>0.000426</td>\n",
       "      <td>0.000760</td>\n",
       "      <td>0.003435</td>\n",
       "      <td>0.000578</td>\n",
       "      <td>0.002797</td>\n",
       "      <td>0.000912</td>\n",
       "      <td>0.001064</td>\n",
       "      <td>0.003618</td>\n",
       "      <td>0.000426</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>falestiniie</th>\n",
       "      <td>0.000742</td>\n",
       "      <td>0.003292</td>\n",
       "      <td>0.001361</td>\n",
       "      <td>0.000569</td>\n",
       "      <td>0.001559</td>\n",
       "      <td>0.000841</td>\n",
       "      <td>0.000272</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.000049</td>\n",
       "      <td>0.000124</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000965</td>\n",
       "      <td>0.001757</td>\n",
       "      <td>0.000742</td>\n",
       "      <td>0.002896</td>\n",
       "      <td>0.001015</td>\n",
       "      <td>0.003242</td>\n",
       "      <td>0.000866</td>\n",
       "      <td>0.001510</td>\n",
       "      <td>0.004405</td>\n",
       "      <td>0.000346</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 194 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             achievement  affection  aggression  air_travel   alcohol  \\\n",
       "crasykayla      0.001277   0.003922    0.000578    0.000304  0.002675   \n",
       "falestiniie     0.000742   0.003292    0.001361    0.000569  0.001559   \n",
       "\n",
       "              ancient     anger    animal  anonymity  anticipation    ...     \\\n",
       "crasykayla   0.000426  0.000334  0.001186   0.000061      0.000061    ...      \n",
       "falestiniie  0.000841  0.000272  0.001138   0.000049      0.000124    ...      \n",
       "\n",
       "              wealthy    weapon   weather   wedding  white_collar_job  \\\n",
       "crasykayla   0.000638  0.000426  0.000760  0.003435          0.000578   \n",
       "falestiniie  0.000965  0.001757  0.000742  0.002896          0.001015   \n",
       "\n",
       "                 work   worship   writing     youth      zest  \n",
       "crasykayla   0.002797  0.000912  0.001064  0.003618  0.000426  \n",
       "falestiniie  0.003242  0.000866  0.001510  0.004405  0.000346  \n",
       "\n",
       "[2 rows x 194 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lexicon_dataframe = pd.DataFrame(list(lexicon_hate_user.values())+list(lexicon_counter_user.values())+list(lexicon_Noncounter_user.values()))\n",
    "lexicon_dataframe.index = list(lexicon_hate_user.keys())+list(lexicon_counter_user.keys())+list(lexicon_Noncounter_user.keys())\n",
    "print(lexicon_dataframe.shape)\n",
    "lexicon_dataframe.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* __lexWords_counter/hate__ are numpy arrays which lexicon contains values corresponding to names_analyized for each counter speaker and hate speaker respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:43:55.199549Z",
     "start_time": "2018-10-29T01:43:55.186844Z"
    }
   },
   "outputs": [],
   "source": [
    "# names_analyized  = ['sadness','fear','swearing_terms','independence','government','love',\n",
    "#                'positive_emotion','negative_emotion','warmth','ridicule','masculine',\n",
    "#                'feminine','violence','suffering','dispute','anger','envy','work','leader',\n",
    "#                'politics','terrorism','shame','confusion','hate']\n",
    "lexWords_counter = lexicon_dataframe.loc[lexicon_counter_user.keys()].values\n",
    "lexWords_hate = lexicon_dataframe.loc[lexicon_hate_user.keys()].values\n",
    "lexWords_Noncounter = lexicon_dataframe.loc[lexicon_Noncounter_user.keys()].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:43:55.726474Z",
     "start_time": "2018-10-29T01:43:55.719393Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.00253187, 0.00088321, 0.00088321, ..., 0.00079489, 0.00288515,\n",
       "        0.00011776],\n",
       "       [0.00171051, 0.00075159, 0.00054425, ..., 0.00142542, 0.00194376,\n",
       "        0.00012958],\n",
       "       [0.00231481, 0.00195738, 0.00175313, ..., 0.00376157, 0.00163399,\n",
       "        0.00027233],\n",
       "       ...,\n",
       "       [0.00162961, 0.00310763, 0.00070743, ..., 0.00131379, 0.00333502,\n",
       "        0.00018949],\n",
       "       [0.001654  , 0.00060434, 0.00114507, ..., 0.00155857, 0.00143134,\n",
       "        0.00012723],\n",
       "       [0.00127798, 0.00058718, 0.00131252, ..., 0.00072534, 0.00079442,\n",
       "        0.00010362]])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lexWords_Noncounter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Checking out p-values for each category :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:44:26.711514Z",
     "start_time": "2018-10-29T01:44:26.684918Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "27"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "useful_categories = np.loadtxt('Useful_Categories.txt',dtype = str)\n",
    "useful_categories = set(useful_categories)-(set(['confusion','divine','emotional','healing'\n",
    "                    ,'kill','joy','love','nervousness','optimism','positive_emotion','suffering'\n",
    "                    ,'appearance','horror','monster','ugliness','zest','worship','shame',\n",
    "                    'affection']))\n",
    "useful_categories = list(useful_categories)\n",
    "len(useful_categories)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:44:28.132370Z",
     "start_time": "2018-10-29T01:44:27.044022Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                      envy\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-2.9609556638828907, pvalue=0.003110338464491982)\n",
      "MannwhitneyuResult(statistic=227380.0, pvalue=8.490165409995618e-14)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-1.4347025955404678, pvalue=0.1518216800245145)\n",
      "MannwhitneyuResult(statistic=33938.0, pvalue=1.8746212022019364e-09)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=0.7512952980023908, pvalue=0.45260678091458717)\n",
      "MannwhitneyuResult(statistic=104105.0, pvalue=0.1826375956634001)\n",
      "\n",
      "                                                      law\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=5.032114058903994, pvalue=5.376920559734352e-07)\n",
      "MannwhitneyuResult(statistic=258884.5, pvalue=5.16095357624846e-05)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=2.6715003230958256, pvalue=0.00772813284359532)\n",
      "MannwhitneyuResult(statistic=39366.0, pvalue=0.00016258855069201754)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=1.2662408232026472, pvalue=0.20564678868707786)\n",
      "MannwhitneyuResult(statistic=104801.0, pvalue=0.2223361755918994)\n",
      "\n",
      "                                                      politics\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=4.498013128715716, pvalue=7.334607180713436e-06)\n",
      "MannwhitneyuResult(statistic=256728.0, pvalue=1.849357653611342e-05)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=1.467281284651615, pvalue=0.14275191551214736)\n",
      "MannwhitneyuResult(statistic=40301.0, pvalue=0.0006869609764228887)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=1.895460168625439, pvalue=0.05824632710887418)\n",
      "MannwhitneyuResult(statistic=108165.0, pvalue=0.4672790052752861)\n",
      "\n",
      "                                                      religion\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=2.386987231435674, pvalue=0.01709832209775008)\n",
      "MannwhitneyuResult(statistic=272795.0, pvalue=0.009576919523028516)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=0.40787587468707487, pvalue=0.6834904206496633)\n",
      "MannwhitneyuResult(statistic=43669.5, pvalue=0.03834085352828626)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=1.2452698754443021, pvalue=0.21325011414920073)\n",
      "MannwhitneyuResult(statistic=107556.5, pvalue=0.41860220145152327)\n",
      "\n",
      "                                                      negative_emotion\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-7.198365040148467, pvalue=9.190546209467217e-13)\n",
      "MannwhitneyuResult(statistic=218400.5, pvalue=3.014490418157234e-17)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-3.7938960128644394, pvalue=0.00016122191360997732)\n",
      "MannwhitneyuResult(statistic=37300.0, pvalue=3.913851200318883e-06)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-0.4615043430149528, pvalue=0.6445116550704016)\n",
      "MannwhitneyuResult(statistic=102940.0, pvalue=0.12682260387420846)\n",
      "\n",
      "                                                      dominant_heirarchical\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=4.379425690357928, pvalue=1.2641903519464234e-05)\n",
      "MannwhitneyuResult(statistic=256063.5, pvalue=1.3459868133967651e-05)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=2.2277306238501007, pvalue=0.02621783999997517)\n",
      "MannwhitneyuResult(statistic=37588.0, pvalue=6.804710921444842e-06)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=0.7874639777580343, pvalue=0.43114946400009113)\n",
      "MannwhitneyuResult(statistic=104048.0, pvalue=0.17938469177158944)\n",
      "\n",
      "                                                      cheerfulness\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-1.979529511960489, pvalue=0.04792118297897662)\n",
      "MannwhitneyuResult(statistic=268551.0, pvalue=0.002460338770541312)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-2.1603171957903493, pvalue=0.03108925333740708)\n",
      "MannwhitneyuResult(statistic=42470.0, pvalue=0.011346753872116715)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=0.9641687133207933, pvalue=0.33513518057076874)\n",
      "MannwhitneyuResult(statistic=106362.0, pvalue=0.3271900882992763)\n",
      "\n",
      "                                                      swearing_terms\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-7.227786800819393, pvalue=7.452757818883202e-13)\n",
      "MannwhitneyuResult(statistic=199718.0, pvalue=8.736703065104208e-26)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-3.197444426780065, pvalue=0.0014493718587284827)\n",
      "MannwhitneyuResult(statistic=36556.0, pvalue=8.540777254664503e-07)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-0.7617815228081768, pvalue=0.4463243609664024)\n",
      "MannwhitneyuResult(statistic=96153.5, pvalue=0.005905870073243574)\n",
      "\n",
      "                                                      ridicule\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-2.285847004005979, pvalue=0.022388579577814267)\n",
      "MannwhitneyuResult(statistic=254268.0, pvalue=5.5744934457269165e-06)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-2.3198079114945367, pvalue=0.020640102598932373)\n",
      "MannwhitneyuResult(statistic=42570.5, pvalue=0.012678480167240162)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=0.42990031284519586, pvalue=0.6673372280093526)\n",
      "MannwhitneyuResult(statistic=105765.5, pvalue=0.28479836344116705)\n",
      "\n",
      "                                                      violence\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-6.180077791865916, pvalue=8.042556446522544e-10)\n",
      "MannwhitneyuResult(statistic=233290.5, pvalue=9.337084480437454e-12)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-2.678524870896303, pvalue=0.007569537986102396)\n",
      "MannwhitneyuResult(statistic=40666.0, pvalue=0.0011706799872766083)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-1.267841279447077, pvalue=0.20507472423446452)\n",
      "MannwhitneyuResult(statistic=101751.5, pvalue=0.08340173338043955)\n",
      "\n",
      "                                                      disappointment\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=2.918477291746657, pvalue=0.0035647547405363053)\n",
      "MannwhitneyuResult(statistic=271169.0, pvalue=0.005775318954704251)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=2.9097090304101174, pvalue=0.003733253796886495)\n",
      "MannwhitneyuResult(statistic=41290.5, pvalue=0.0027108067517563356)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-1.4931080997275468, pvalue=0.1356441228466919)\n",
      "MannwhitneyuResult(statistic=103533.5, pvalue=0.15326558330023804)\n",
      "\n",
      "                                                      hate\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-3.715955323817305, pvalue=0.00020913848734479195)\n",
      "MannwhitneyuResult(statistic=236990.5, pvalue=1.428493346727184e-10)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-2.2209031791509077, pvalue=0.02667911127955518)\n",
      "MannwhitneyuResult(statistic=37636.0, pvalue=7.543051685900367e-06)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=0.9610615921140335, pvalue=0.33669445325734404)\n",
      "MannwhitneyuResult(statistic=108389.0, pvalue=0.4854013279999447)\n",
      "\n",
      "                                                      masculine\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-6.4320485198641455, pvalue=1.6416696804046177e-10)\n",
      "MannwhitneyuResult(statistic=227656.0, pvalue=1.0681875466263632e-13)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-1.3077100664562449, pvalue=0.19140406666847878)\n",
      "MannwhitneyuResult(statistic=45353.0, pvalue=0.14537457931933256)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-3.04001763955678, pvalue=0.0024114729127860516)\n",
      "MannwhitneyuResult(statistic=89168.5, pvalue=4.1741102078267164e-05)\n",
      "\n",
      "                                                      death\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-3.825868027572278, pvalue=0.00013513311450579285)\n",
      "MannwhitneyuResult(statistic=244214.0, pvalue=1.829657721971377e-08)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-2.4374192129860757, pvalue=0.015042253293738799)\n",
      "MannwhitneyuResult(statistic=40756.5, pvalue=0.0013289246732429974)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-0.10659737019940022, pvalue=0.9151243489487938)\n",
      "MannwhitneyuResult(statistic=104572.0, pvalue=0.20879518382347495)\n",
      "\n",
      "                                                      dominant_personality\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-3.3160050392910922, pvalue=0.0009328189879789349)\n",
      "MannwhitneyuResult(statistic=251671.0, pvalue=1.4216571639272366e-06)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-3.496415827134167, pvalue=0.0005015352598434606)\n",
      "MannwhitneyuResult(statistic=40800.0, pvalue=0.0014107365727356422)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=1.1322812681497016, pvalue=0.257718621415641)\n",
      "MannwhitneyuResult(statistic=107370.5, pvalue=0.4039274819984755)\n",
      "\n",
      "                                                      terrorism\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=2.462015535248912, pvalue=0.013916741641138987)\n",
      "MannwhitneyuResult(statistic=269304.5, pvalue=0.003154722972728662)\n",
      "Hate-----NonCounter\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ttest_indResult(statistic=0.38941789158605217, pvalue=0.6970863312017319)\n",
      "MannwhitneyuResult(statistic=41686.5, pvalue=0.004498962899892931)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=1.3016270887783485, pvalue=0.1932674528326792)\n",
      "MannwhitneyuResult(statistic=106023.0, pvalue=0.302610111338193)\n",
      "\n",
      "                                                      politeness\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-0.6099349200808981, pvalue=0.5419882998457117)\n",
      "MannwhitneyuResult(statistic=271824.0, pvalue=0.007146776088372802)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=0.5022809713223286, pvalue=0.615628990763603)\n",
      "MannwhitneyuResult(statistic=47335.5, pvalue=0.414424615703828)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-0.7424720371081706, pvalue=0.457931419013598)\n",
      "MannwhitneyuResult(statistic=99273.0, pvalue=0.029686579925602065)\n",
      "\n",
      "                                                      anger\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=3.478151047906045, pvalue=0.0005178973212783461)\n",
      "MannwhitneyuResult(statistic=267759.0, pvalue=0.0018598680486770739)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=3.8469659869974313, pvalue=0.00013058857924471204)\n",
      "MannwhitneyuResult(statistic=40291.0, pvalue=0.0006800371541608873)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-2.0180377494104054, pvalue=0.04378614134553031)\n",
      "MannwhitneyuResult(statistic=102445.0, pvalue=0.1070176427015383)\n",
      "\n",
      "                                                      pain\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-2.9846169911773974, pvalue=0.00288077895313071)\n",
      "MannwhitneyuResult(statistic=239824.5, pvalue=1.0326874424413318e-09)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-1.8987465700722603, pvalue=0.05801171647940428)\n",
      "MannwhitneyuResult(statistic=41732.0, pvalue=0.00478085688476621)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-0.5766963840343167, pvalue=0.5642413629704353)\n",
      "MannwhitneyuResult(statistic=101282.5, pvalue=0.06975043117549633)\n",
      "\n",
      "                                                      feminine\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-3.8865551061385633, pvalue=0.00010567216758366788)\n",
      "MannwhitneyuResult(statistic=242010.5, pvalue=4.446702090482048e-09)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-2.80632596943173, pvalue=0.005151270526349727)\n",
      "MannwhitneyuResult(statistic=41105.0, pvalue=0.0021378807316763727)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-0.08311577604015681, pvalue=0.9337718864147125)\n",
      "MannwhitneyuResult(statistic=103781.5, pvalue=0.16579467649233098)\n",
      "\n",
      "                                                      independence\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=2.399824127593705, pvalue=0.01651240501934472)\n",
      "MannwhitneyuResult(statistic=268983.5, pvalue=0.002807835296762422)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=3.4881113717339987, pvalue=0.0005170946288445249)\n",
      "MannwhitneyuResult(statistic=39495.5, pvalue=0.00019656805458890405)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-0.34785760551268136, pvalue=0.7280017184349055)\n",
      "MannwhitneyuResult(statistic=100870.5, pvalue=0.05886169288287138)\n",
      "\n",
      "                                                      government\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=5.121931866769678, pvalue=3.3764145908235574e-07)\n",
      "MannwhitneyuResult(statistic=255106.0, pvalue=8.471681661033113e-06)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=2.375691026021756, pvalue=0.01778630857782219)\n",
      "MannwhitneyuResult(statistic=38885.5, pvalue=7.28499369540312e-05)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=1.430302355393579, pvalue=0.1528632122852728)\n",
      "MannwhitneyuResult(statistic=105294.5, pvalue=0.2532520497955393)\n",
      "\n",
      "                                                      torment\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=3.135467064671099, pvalue=0.0017458363102698848)\n",
      "MannwhitneyuResult(statistic=271391.5, pvalue=0.00624409348954226)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=3.4209479789452635, pvalue=0.0006605431036877892)\n",
      "MannwhitneyuResult(statistic=42225.0, pvalue=0.008597470548804127)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-1.8552661883694974, pvalue=0.0637775407793741)\n",
      "MannwhitneyuResult(statistic=104246.0, pvalue=0.19025268906861986)\n",
      "\n",
      "                                                      leader\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=3.8020115095797617, pvalue=0.00014871049252026644)\n",
      "MannwhitneyuResult(statistic=268249.0, pvalue=0.0022168421417042943)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=1.0379053181161793, pvalue=0.29967509810596626)\n",
      "MannwhitneyuResult(statistic=41840.5, pvalue=0.005458804900294265)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=1.6788043528612138, pvalue=0.0934231537548242)\n",
      "MannwhitneyuResult(statistic=106995.0, pvalue=0.3747479156157526)\n",
      "\n",
      "                                                      sexual\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-4.052773765633726, pvalue=5.295059406668528e-05)\n",
      "MannwhitneyuResult(statistic=245549.5, pvalue=4.183985459600581e-08)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=-0.5688794801416436, pvalue=0.5696217136310497)\n",
      "MannwhitneyuResult(statistic=45295.0, pvalue=0.13983158370765159)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-2.0691535174353977, pvalue=0.03872289282802049)\n",
      "MannwhitneyuResult(statistic=96589.5, pvalue=0.007559972015881457)\n",
      "\n",
      "                                                      disgust\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=3.5613202094383145, pvalue=0.00037939095115281695)\n",
      "MannwhitneyuResult(statistic=269288.0, pvalue=0.003135324359488212)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=3.5488831474956473, pvalue=0.00041290933832878446)\n",
      "MannwhitneyuResult(statistic=40066.0, pvalue=0.0004847045112175266)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=-1.0737255495169993, pvalue=0.2831388746883018)\n",
      "MannwhitneyuResult(statistic=101804.5, pvalue=0.08483934164398399)\n",
      "\n",
      "                                                      pride\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=2.4863542663503684, pvalue=0.013003502711808513)\n",
      "MannwhitneyuResult(statistic=271037.0, pvalue=0.005587313311768384)\n",
      "Hate-----NonCounter\n",
      "Ttest_indResult(statistic=0.24437194023086217, pvalue=0.8070148968192897)\n",
      "MannwhitneyuResult(statistic=47308.0, pvalue=0.40988352797689914)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=1.3136327131506638, pvalue=0.1891945286454437)\n",
      "MannwhitneyuResult(statistic=99257.0, pvalue=0.02944726572475309)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for j in useful_categories:\n",
    "    _,pvalue1 = ttest_ind(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue2 = mannwhitneyu(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue3 = ttest_ind(lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue4 = mannwhitneyu(lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue5 = ttest_ind(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j])\n",
    "    _,pvalue6 = mannwhitneyu(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j])\n",
    "    print(\"                                                     \",j)\n",
    "    print('Hate----Counter')\n",
    "    print(ttest_ind(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j]))\n",
    "    print(mannwhitneyu(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j]))\n",
    "    print('Hate-----NonCounter')\n",
    "    print(ttest_ind(lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j]))\n",
    "    print(mannwhitneyu(lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j]))\n",
    "    print()\n",
    "    print('Counter----NonCounter')\n",
    "    print(ttest_ind(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "    print(mannwhitneyu(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-29T01:44:35.926736Z",
     "start_time": "2018-10-29T01:44:31.662887Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/navish/anaconda3/lib/python3.6/site-packages/matplotlib/pyplot.py:537: RuntimeWarning: More than 20 figures have been opened. Figures created through the pyplot interface (`matplotlib.pyplot.figure`) are retained until explicitly closed and may consume too much memory. (To control this warning, see the rcParam `figure.max_open_warning`).\n",
      "  max_open_warning, RuntimeWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x504 with 27 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 158.4x144 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "sns.set(palette = 'deep')\n",
    "# for i,j in enumerate(list(names_analyized)):\n",
    "f,axes = plt.subplots(3,9,figsize = (16,7))\n",
    "row,col = 0,0\n",
    "for j in useful_categories:\n",
    "#     _,pvalue1 = ttest_ind(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j])\n",
    "#     _,pvalue2 = mannwhitneyu(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j])\n",
    "#      _,pvalue3 = ttest_ind(lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j])\n",
    "#     _,pvalue4 = mannwhitneyu(lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j])\n",
    "#     _,pvalue5 = ttest_ind(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j])\n",
    "#     _,pvalue6 = mannwhitneyu(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j])\n",
    "#     if  pvalue2 < 0.01 or pvalue4 < 0.01 or pvalue6 < 0.01:\n",
    "#             useful_categories.append(j)\n",
    "#             print(\"                                                     \",j)\n",
    "#             print('Hate----Counter')\n",
    "#             print(ttest_ind(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j]))\n",
    "#             print(mannwhitneyu(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j]))\n",
    "# #             print('           Hate Vs NonCounter')\n",
    "# #             print(ttest_ind(lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j]))\n",
    "# #             print(mannwhitneyu(lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j],lexicon_dataframe.loc[lexicon_hate_user.keys()][j]))\n",
    "#             print()\n",
    "#             print('Counter----NonCounter')\n",
    "#             print(ttest_ind(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "#             print(mannwhitneyu(lexicon_dataframe.loc[lexicon_counter_user.keys()][j],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "#             print()\n",
    "            Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(lexicon_Noncounter_user.keys()))],lexicon_dataframe.loc[lexicon_Noncounter_user.keys()][j].tolist()],columns = ['Users',''])\n",
    "            counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(lexicon_counter_user.keys()))],lexicon_dataframe.loc[lexicon_counter_user.keys()][j].tolist()],columns = ['Users',''])\n",
    "            hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(lexicon_hate_user.keys()))],lexicon_dataframe.loc[lexicon_hate_user.keys()][j].tolist()],columns = ['Users',''])\n",
    "            Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "            Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "#             palette = {'Counter':'midnightblue','Hate':'crimson','NonCounter':'forestgreen'}\n",
    "            plt.figure(figsize=(2.2,2))\n",
    "            sns.barplot(x = 'Users',y ='',data = Comb_df ,errwidth = 1.2,ax = axes[row,col],palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "            axes[row,col].set_xticklabels('')\n",
    "            axes[row,col].set_xlabel('')\n",
    "            if '_' in re.findall(r'\\w',j):\n",
    "                if len(re.findall('[^_]+',j)[0])>3 and re.findall(r'[^_]+',j)[1]!='terms':\n",
    "                    axes[row,col].set_title(re.findall(r'[^_]+',j)[0][0].upper()+re.findall(r'[^_]+',j)[0][1:3]+'.'+' '+re.findall(r'[^_]+',j)[1])\n",
    "                else:\n",
    "                    if re.findall(r'[^_]+',j)[1]=='terms':\n",
    "                        axes[row,col].set_title(re.findall(r'[^_]+',j)[0][0].upper()+re.findall(r'[^_]+',j)[0][1:])\n",
    "                    else:\n",
    "                        print(re.findall(r'[^_]+',j)[1])\n",
    "                        axes[row,col].set_title(re.findall(r'[^_]+',j)[0][0].upper()+re.findall(r'[^_]+',j)[0][1:]+' '+re.findall(r'[^_]+',j)[1]) \n",
    "            else:\n",
    "                axes[row,col].set_title(j[0].upper()+j[1:])\n",
    "            col+=1\n",
    "            if col==9:\n",
    "                row+=1\n",
    "                col=0\n",
    "            if row==3:\n",
    "                break\n",
    "#             def change_width(ax, new_value) :\n",
    "#                 for patch in ax.patches :\n",
    "#                     current_width = patch.get_width()\n",
    "#                     diff = current_width - new_value\n",
    "\n",
    "#                     # we change the bar width\n",
    "#                     patch.set_width(new_value)\n",
    "\n",
    "#                     # we recenter the bar\n",
    "#                     patch.set_x(patch.get_x() + diff * .5)\n",
    "\n",
    "#             change_width(ax, .4)\n",
    "#             plt.savefig('Images/empath_'+ j +'.eps', format='eps') \n",
    "f.subplots_adjust(top=10)\n",
    "import matplotlib.patches as mpatches\n",
    "import matplotlib.pyplot as plt\n",
    "red_patch = mpatches.Patch(color='#c44e52', label='Hate')\n",
    "blue_patch = mpatches.Patch(color='#4c72b0', label='NonCounter')\n",
    "green_patch = mpatches.Patch(color='#55a868', label='Counter')\n",
    "f.legend(handles=[green_patch,red_patch],loc=9,ncol=3, frameon=False, borderaxespad =-0.45)\n",
    "\n",
    "f.tight_layout()\n",
    "f.savefig('Images/empath.eps',format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating a Sentiment DataFrame/TextBlob DataFrame for all Users : "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:12.398989Z",
     "start_time": "2018-10-23T21:40:12.370828Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1852, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>compound</th>\n",
       "      <th>neg</th>\n",
       "      <th>neu</th>\n",
       "      <th>pos</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>crasykayla</th>\n",
       "      <td>0.060108</td>\n",
       "      <td>0.110590</td>\n",
       "      <td>0.740262</td>\n",
       "      <td>0.149144</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>falestiniie</th>\n",
       "      <td>0.011887</td>\n",
       "      <td>0.095849</td>\n",
       "      <td>0.792552</td>\n",
       "      <td>0.111597</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             compound       neg       neu       pos\n",
       "crasykayla   0.060108  0.110590  0.740262  0.149144\n",
       "falestiniie  0.011887  0.095849  0.792552  0.111597"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sentiment_dataframe = pd.DataFrame(list(sentiment_hate_user.values())+list(sentiment_counter_user.values())+list(sentiment_Noncounter_user.values()))\n",
    "sentiment_dataframe.index = list(sentiment_hate_user.keys())+list(sentiment_counter_user.keys())+list(sentiment_Noncounter_user.keys())\n",
    "print(sentiment_dataframe.shape)\n",
    "sentiment_dataframe.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:12.656327Z",
     "start_time": "2018-10-23T21:40:12.649897Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'polarity': 0.11299070730567451, 'subjectivity': 0.5946555603785325}"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(textblob_hate_user.values())[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:12.906325Z",
     "start_time": "2018-10-23T21:40:12.883954Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1852, 2)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>polarity</th>\n",
       "      <th>subjectivity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>crasykayla</th>\n",
       "      <td>0.112991</td>\n",
       "      <td>0.594656</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>falestiniie</th>\n",
       "      <td>0.044500</td>\n",
       "      <td>0.542153</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             polarity  subjectivity\n",
       "crasykayla   0.112991      0.594656\n",
       "falestiniie  0.044500      0.542153"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "textblob_dataframe = pd.DataFrame(list(textblob_hate_user.values())+list(textblob_counter_user.values())+list(textblob_Noncounter_user.values()))\n",
    "textblob_dataframe.index = list(textblob_hate_user.keys())+list(textblob_counter_user.keys())+list(textblob_Noncounter_user.keys())\n",
    "print(textblob_dataframe.shape)\n",
    "textblob_dataframe.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* __senWords_counter/hate__ are numpy arrays which contains sentiment values corresponding to each user."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:13.200542Z",
     "start_time": "2018-10-23T21:40:13.183328Z"
    }
   },
   "outputs": [],
   "source": [
    "senWords_counter = sentiment_dataframe.loc[sentiment_counter_user.keys()].values\n",
    "senWords_hate = sentiment_dataframe.loc[sentiment_hate_user.keys()].values\n",
    "senWords_Noncounter = sentiment_dataframe.loc[sentiment_Noncounter_user.keys()].values\n",
    "#--------------------------------------------------------------------------------------------------\n",
    "textblob_counter = textblob_dataframe.loc[textblob_counter_user.keys()].values\n",
    "textblob_hate = textblob_dataframe.loc[textblob_hate_user.keys()].values\n",
    "textblob_Noncounter = textblob_dataframe.loc[textblob_Noncounter_user.keys()].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:13.378603Z",
     "start_time": "2018-10-23T21:40:13.372028Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.30403886, 0.59109541],\n",
       "       [0.10960933, 0.54541847],\n",
       "       [0.16809592, 0.58320833],\n",
       "       ...,\n",
       "       [0.07356571, 0.57161889],\n",
       "       [0.14583333, 0.46666667],\n",
       "       [0.0075    , 0.50333333]])"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# senWords_counter\n",
    "textblob_counter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sentiment Plotting "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### T-Test on Sentiment Values:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T22:31:41.564653Z",
     "start_time": "2018-10-23T22:31:41.038675Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x144 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(palette = 'deep')\n",
    "f,axes = plt.subplots(1,3,figsize = (4,2))\n",
    "col = 0\n",
    "for i,j in enumerate(['Compound','Negative','Neutral','Positive']):\n",
    "        _,pvalue1 = mannwhitneyu(senWords_counter[:,i],senWords_hate[:,i])\n",
    "        _,pvalue2 = mannwhitneyu(senWords_Noncounter[:,i],senWords_hate[:,i])\n",
    "        _,pvalue3 = mannwhitneyu(senWords_counter[:,i],senWords_Noncounter[:,i])\n",
    "        if pvalue1<0.01 or pvalue2<0.01 or pvalue3<0.01:\n",
    "#             print('                                      ',j)\n",
    "#             print('Hate----Counter')\n",
    "#             print(ttest_ind(senWords_counter[:,i],senWords_hate[:,i]))\n",
    "#             print(mannwhitneyu(senWords_counter[:,i],senWords_hate[:,i]))\n",
    "#             print()\n",
    "#             print('Counter----Non Counter')\n",
    "#             print(ttest_ind(senWords_counter[:,i],senWords_Noncounter[:,i]))\n",
    "#             print(mannwhitneyu(senWords_counter[:,i],senWords_Noncounter[:,i]))\n",
    "#             print()\n",
    "#             print('Hate----Non Counter')\n",
    "#             print(ttest_ind(senWords_hate[:,i],senWords_Noncounter[:,i]))\n",
    "#             print(mannwhitneyu(senWords_hate[:,i],senWords_Noncounter[:,i]))\n",
    "            Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(senWords_Noncounter[:,i]))],senWords_Noncounter[:,i]],columns = ['Users',''])\n",
    "            counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(senWords_counter[:,i]))],senWords_counter[:,i]],columns = ['Users',''])\n",
    "            hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(senWords_hate[:,i]))],senWords_hate[:,i]],columns = ['Users',''])\n",
    "            Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "            Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "#             sns.boxplot(x = 'Users',y ='',data = Comb_df, linewidth=1,width=0.6,showfliers = False,ax = axes[col])\n",
    "#             axes[col].set_xticklabels('')\n",
    "#             axes[col].set_xlabel('')\n",
    "#             axes[col].set_title(j[0].upper()+j[1:])\n",
    "# #             col+=1\n",
    "# f.tight_layout()\n",
    "# f.savefig('Images/sentiment_Boxplots.eps', format='eps')\n",
    "# plt.show()\n",
    "######################################### Barplot##################################### \n",
    "            sns.barplot(x = 'Users',y ='',data = Comb_df,errwidth=1.2,ax =axes[col],palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "            axes[col].set_xticklabels('')\n",
    "            axes[col].set_xlabel('')\n",
    "            axes[col].set_title(j[0].upper()+j[1:])\n",
    "            col+=1\n",
    "            \n",
    "f.tight_layout()\n",
    "f.savefig('Images/sentiment_Barplots.eps', format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T22:31:53.709237Z",
     "start_time": "2018-10-23T22:31:53.535364Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 158.4x144 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(senWords_Noncounter[:,1]))],senWords_Noncounter[:,1]],columns = ['Users',''])\n",
    "counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(senWords_counter[:,1]))],senWords_counter[:,1]],columns = ['Users',''])\n",
    "hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(senWords_hate[:,1]))],senWords_hate[:,1]],columns = ['Users',''])\n",
    "Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "plt.figure(figsize = (2.2,2))\n",
    "ax  = sns.boxplot(x = 'Users',y ='',data = Comb_df, linewidth=1,width=0.6,showfliers = False,palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "ax.set_xticklabels('')\n",
    "ax.set_xlabel('')\n",
    "ax.set_title('Sentiment')\n",
    "plt.tight_layout()\n",
    "plt.savefig('Images/Negative.eps', format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T22:32:39.966192Z",
     "start_time": "2018-10-23T22:32:39.529256Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                       polarity\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=2.455632421009419, pvalue=0.01416543976209231)\n",
      "MannwhitneyuResult(statistic=280752.0, pvalue=0.07201104403840128)\n",
      "\n",
      "Counter----Non Counter\n",
      "Ttest_indResult(statistic=-0.10917902957438345, pvalue=0.9130768274045733)\n",
      "MannwhitneyuResult(statistic=105050.0, pvalue=0.23771329166729888)\n",
      "\n",
      "Hate----Non Counter\n",
      "Ttest_indResult(statistic=-1.8310454896268553, pvalue=0.06752171728197422)\n",
      "MannwhitneyuResult(statistic=44008.0, pvalue=0.05191393849464162)\n",
      "                                       subjectivity\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=-6.542568233560775, pvalue=8.029660681731519e-11)\n",
      "MannwhitneyuResult(statistic=235308.0, pvalue=4.21773148920915e-11)\n",
      "\n",
      "Counter----Non Counter\n",
      "Ttest_indResult(statistic=0.45801599519250924, pvalue=0.6470151553258225)\n",
      "MannwhitneyuResult(statistic=105192.0, pvalue=0.24670885296358674)\n",
      "\n",
      "Hate----Non Counter\n",
      "Ttest_indResult(statistic=4.12135249781638, pvalue=4.22115463931937e-05)\n",
      "MannwhitneyuResult(statistic=35854.0, pvalue=1.8600621788880612e-07)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x180 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 158.4x144 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f,axes = plt.subplots(1,2,figsize = (5,2.5))\n",
    "col = 0\n",
    "for i,j in enumerate(['polarity','subjectivity']):\n",
    "        _,pvalue1 = mannwhitneyu(textblob_counter[:,i],textblob_hate[:,i])\n",
    "        _,pvalue2 = mannwhitneyu(textblob_Noncounter[:,i],textblob_hate[:,i])\n",
    "        _,pvalue3 = mannwhitneyu(textblob_counter[:,i],textblob_Noncounter[:,i])\n",
    "        if pvalue1<0.1 or pvalue2<0.1 or pvalue3<0.1:\n",
    "            print('                                      ',j)\n",
    "            print('Hate----Counter')\n",
    "            print(ttest_ind(senWords_counter[:,i],senWords_hate[:,i]))\n",
    "            print(mannwhitneyu(senWords_counter[:,i],senWords_hate[:,i]))\n",
    "            print()\n",
    "            print('Counter----Non Counter')\n",
    "            print(ttest_ind(senWords_counter[:,i],senWords_Noncounter[:,i]))\n",
    "            print(mannwhitneyu(senWords_counter[:,i],senWords_Noncounter[:,i]))\n",
    "            print()\n",
    "            print('Hate----Non Counter')\n",
    "            print(ttest_ind(senWords_hate[:,i],senWords_Noncounter[:,i]))\n",
    "            print(mannwhitneyu(senWords_hate[:,i],senWords_Noncounter[:,i]))\n",
    "            Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(textblob_Noncounter[:,i]))],textblob_Noncounter[:,i]],columns = ['Users',''])\n",
    "            counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(textblob_counter[:,i]))],textblob_counter[:,i]],columns = ['Users',''])\n",
    "            hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(textblob_hate[:,i]))],textblob_hate[:,i]],columns = ['Users',''])\n",
    "            Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "            Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "            sns.boxplot(x = 'Users',y ='',data = Comb_df, linewidth=1,width=0.6,showfliers = False,ax = axes[col],palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "            axes[col].set_xticklabels('')\n",
    "            axes[col].set_xlabel('')\n",
    "            axes[col].set_title(j[0].upper()+j[1:])\n",
    "            col+=1\n",
    "f.tight_layout()\n",
    "f.savefig('Images/TextBlob_Boxplots.eps', format='eps')\n",
    "plt.show()\n",
    "######################################### Barplot##################################### \n",
    "#             sns.barplot(x = 'Users',y ='',data = Comb_df,errwidth=1.2,ax =axes[col])\n",
    "#             axes[col].set_xticklabels('')\n",
    "#             axes[col].set_xlabel('')\n",
    "#             axes[col].set_title(j[0].upper()+j[1:])\n",
    "#             col+=1\n",
    "# f.tight_layout()\n",
    "# f.savefig('Images/sentiment_Barplots', format='eps')\n",
    "# plt.show()\n",
    "Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(textblob_Noncounter[:,1]))],textblob_Noncounter[:,1]],columns = ['Users',''])\n",
    "counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(textblob_counter[:,1]))],textblob_counter[:,1]],columns = ['Users',''])\n",
    "hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(textblob_hate[:,1]))],textblob_hate[:,1]],columns = ['Users',''])\n",
    "Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "plt.figure(figsize = (2.2,2))\n",
    "ax = sns.boxplot(x = 'Users',y ='',data = Comb_df, linewidth=1,width=0.6,showfliers = False,palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "ax.set_xticklabels('')\n",
    "ax.set_xlabel('')\n",
    "ax.set_title('Subjectivity')\n",
    "plt.tight_layout()\n",
    "plt.savefig('Images/Subjectivity.eps', format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* user_tweetid.pkl :: Contains tweet ids coressponding to each user name that it has tweeted"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:20.026443Z",
     "start_time": "2018-10-23T21:40:18.512437Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1852"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open('Refined_Project_Dataset/user_tweetid.pkl','rb') as f:\n",
    "    names = pickle.load(f).keys()\n",
    "len(names) ##### user names foundout during crawling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### missing_values either due to account deletion or protected mode or warning time not exceeded."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:21.353610Z",
     "start_time": "2018-10-23T21:40:21.345084Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1972\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "120"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(len(all_usernames))\n",
    "missing_values  = [i for i in set(all_usernames)-set(names)]\n",
    "len(missing_values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:21.555710Z",
     "start_time": "2018-10-23T21:40:21.548574Z"
    }
   },
   "outputs": [],
   "source": [
    "# with open('Refined_Project_Dataset/missing_users_binny.txt','w') as fp:\n",
    "temp_missing = [i for i in set(all_usernames)-set(names)]\n",
    "#     print(len(temp_missing))\n",
    "#     for i in temp_missing:\n",
    "#         fp.write(i)\n",
    "#         fp.write('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:21.704348Z",
     "start_time": "2018-10-23T21:40:21.697573Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(66, 39, 16)"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "##### Distribution of missing users among hate , counter and Non-Counter.\n",
    "len(set(temp_missing).intersection(counter_usernames)), len(set(temp_missing).intersection(hate_usernames)),len(set(temp_missing).intersection(Non_counter_usernames))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Finding out to which category missing counterspeech belong:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:22.511082Z",
     "start_time": "2018-10-23T21:40:22.364919Z"
    }
   },
   "outputs": [],
   "source": [
    "missing_counter_labels = []\n",
    "missing_hate_labels = []\n",
    "missing_Noncounter_labels = []\n",
    "for i in counter_usernames:\n",
    "    if i in temp_missing:\n",
    "        missing_counter_labels.append(df[df['replier_user_name']== i]['tag'].values[0])\n",
    "\n",
    "for i in hate_usernames:\n",
    "    if i in temp_missing:\n",
    "        missing_hate_labels.append(df[df['original_tweetusername']== i]['tag'].values[0])\n",
    "\n",
    "for i in Non_counter_usernames:\n",
    "    if i in temp_missing:\n",
    "        missing_Noncounter_labels.append(df[df['replier_user_name']== i]['tag'].values[0])\n",
    "# np.savetxt('missing_counter_labels.txt',missing_counter_labels,fmt = '%s') \n",
    "# np.savetxt('missing_hate_labels.txt',missing_hate_labels,fmt = '%s') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:23.087969Z",
     "start_time": "2018-10-23T21:40:23.081305Z"
    }
   },
   "outputs": [],
   "source": [
    "categories = ['0','1','2','3','4','5','6','7','8','9','11']\n",
    "dict_counter = {}\n",
    "dict_hate = {}\n",
    "for i in categories:\n",
    "    dict_counter[i] = (re.findall(r'\\d+',(' '.join(missing_counter_labels)))).count(i)\n",
    "                \n",
    "for i in categories:\n",
    "    dict_hate[i] = (re.findall(r'\\d+',(' '.join(missing_hate_labels)))).count(i)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Distribution of __categories in missing counter speeches__:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:23.437972Z",
     "start_time": "2018-10-23T21:40:23.431944Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'0': 0,\n",
       " '1': 7,\n",
       " '2': 5,\n",
       " '3': 6,\n",
       " '4': 2,\n",
       " '5': 11,\n",
       " '6': 15,\n",
       " '7': 8,\n",
       " '8': 8,\n",
       " '9': 9,\n",
       " '11': 8}"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dict_counter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:23.744373Z",
     "start_time": "2018-10-23T21:40:23.738453Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'0': 7,\n",
       " '1': 1,\n",
       " '2': 3,\n",
       " '3': 3,\n",
       " '4': 2,\n",
       " '5': 4,\n",
       " '6': 5,\n",
       " '7': 2,\n",
       " '8': 5,\n",
       " '9': 10,\n",
       " '11': 4}"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dict_hate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:24.086243Z",
     "start_time": "2018-10-23T21:40:24.079260Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(120, 548, 1239, 220)"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(set(temp_missing)),len(set(hate_usernames)),len(counter_usernames),len(Non_counter_usernames)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## % Missing Hate speakers and counter speakers:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:24.678341Z",
     "start_time": "2018-10-23T21:40:24.670985Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of Hatespeech missing 0.07116788321167883\n",
      "Percentage of Counterspeech missing 0.053268765133171914\n",
      "Percentage of Non-Counterspeech missing 0.07272727272727272\n"
     ]
    }
   ],
   "source": [
    "print('Percentage of Hatespeech missing',39/548)\n",
    "print('Percentage of Counterspeech missing',66/1239)\n",
    "print('Percentage of Non-Counterspeech missing',16/220)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### user_info_characteristics.json : Contains all user information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:26.877302Z",
     "start_time": "2018-10-23T21:40:25.422825Z"
    }
   },
   "outputs": [],
   "source": [
    "with open('Refined_Project_Dataset/user_info.pkl','rb') as fp:\n",
    "    data = pickle.load(fp)\n",
    "# user_info_characteristics = []\n",
    "# for value,tweets in zip(data.values(),all_user_tweets.values()):\n",
    "#     value.update({'num_tweets':len(tweets)})\n",
    "#     user_info_characteristics.append(value)\n",
    "# with open('Refined_Project_Dataset/user_info_characteristics.json','w') as fp:\n",
    "#     json.dump(user_info_characteristics,fp)\n",
    "    \n",
    "## Creating DataFrame : \n",
    "user_info_characteristics = pd.read_json('Refined_Project_Dataset/user_info_characteristics.json')\n",
    "user_info_characteristics.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Finding out the total count / list of Followers...  for all the users combined"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:30.758870Z",
     "start_time": "2018-10-23T21:40:29.894552Z"
    }
   },
   "outputs": [],
   "source": [
    "followers_count_Hate_list,friends_count_Hate_list,listed_count_Hate_list,favourites_count_Hate_list,statuses_count_Hate_list = [],[],[],[],[]\n",
    "followers_count_Counter_list,friends_count_Counter_list,listed_count_Counter_list,statuses_count_Counter_list,favourites_count_Counter_list = [],[],[],[],[]\n",
    "followers_count_NonCounter_list,friends_count_NonCounter_list,listed_count_NonCounter_list,statuses_count_NonCounter_list,favourites_count_NonCounter_list = [],[],[],[],[]\n",
    "\n",
    "for key in names:\n",
    "    if key.lower().strip() in [i.lower().strip() for i in hate_usernames]:\n",
    "        followers_count_Hate_list.append(data[key]['followers_count'])\n",
    "        friends_count_Hate_list.append(data[key]['friends_count'])\n",
    "        listed_count_Hate_list.append(data[key]['listed_count'])\n",
    "        favourites_count_Hate_list.append(data[key]['favourites_count'])\n",
    "        statuses_count_Hate_list.append(data[key]['statuses_count'])\n",
    "    \n",
    "for key in names:\n",
    "    if key.lower().strip() in [i.lower().strip() for i in counter_usernames]:\n",
    "        followers_count_Counter_list.append(data[key]['followers_count'])\n",
    "        friends_count_Counter_list.append(data[key]['friends_count'])\n",
    "        listed_count_Counter_list.append(data[key]['listed_count'])\n",
    "        favourites_count_Counter_list.append(data[key]['favourites_count'])\n",
    "        statuses_count_Counter_list.append(data[key]['statuses_count'])\n",
    "\n",
    "            \n",
    "for key in names:\n",
    "    if key.lower().strip() in [i.lower().strip() for i in Non_counter_usernames]:\n",
    "        followers_count_NonCounter_list.append(data[key]['followers_count'])\n",
    "        friends_count_NonCounter_list.append(data[key]['friends_count'])\n",
    "        listed_count_NonCounter_list.append(data[key]['listed_count'])\n",
    "        favourites_count_NonCounter_list.append(data[key]['favourites_count'])\n",
    "        statuses_count_NonCounter_list.append(data[key]['statuses_count'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:30.762814Z",
     "start_time": "2018-10-23T21:40:30.760478Z"
    }
   },
   "outputs": [],
   "source": [
    "#####CAUTION : Please do not run this as such,first do required changes in above function \n",
    "# print(followers_count_Hate,friends_count_Hate,listed_count_Hate,favourites_count_Hate,statuses_count_Hate)\n",
    "# print(followers_count_Counter,friends_count_Counter,listed_count_Counter,favourites_count_Counter,statuses_count_Counter)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Total no of intersection points in craawled data and the original data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:31.656787Z",
     "start_time": "2018-10-23T21:40:31.645252Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1852"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## length of common elements without lowercasing for checking:\n",
    "len(set([i.lower().strip() for i in data.keys()]).intersection(set([i.lower().strip() for i in list(hate_usernames)+list(counter_usernames)+list(Non_counter_usernames)])))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# T-Test :"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Modying features of DataFrame to get numerical data: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:34.182883Z",
     "start_time": "2018-10-23T21:40:34.110326Z"
    }
   },
   "outputs": [],
   "source": [
    "#### Getting only dates from creation date and time column.\n",
    "def get_dates(d):\n",
    "    d = str(d)\n",
    "    date = []\n",
    "    date.append(re.findall(r'\\w\\w\\w\\w-\\w+-\\w\\w',d)[0])\n",
    "    return date\n",
    "user_info_characteristics['date'] = user_info_characteristics['created_at'].apply(get_dates)\n",
    "\n",
    "## modyfying features :\n",
    "\n",
    "user_info_characteristics['default_profile'] = user_info_characteristics['default_profile'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['default_profile_image'] = user_info_characteristics['default_profile_image'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['verified'] = user_info_characteristics['verified'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['protected'] = user_info_characteristics['protected'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['geo_enabled'] = user_info_characteristics['geo_enabled'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['contributors_enabled'] = user_info_characteristics['contributors_enabled'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['is_translator'] = user_info_characteristics['is_translator'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['is_translation_enabled'] = user_info_characteristics['is_translation_enabled'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['profile_background_tile'] = user_info_characteristics['profile_background_tile'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['profile_use_background_image'] = user_info_characteristics['profile_use_background_image'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['has_extended_profile'] = user_info_characteristics['has_extended_profile'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['default_profile'] = user_info_characteristics['default_profile'].apply(lambda x: 1 if x ==True else 0)\n",
    "user_info_characteristics['default_profile_image'] = user_info_characteristics['default_profile_image'].apply(lambda x: 1 if x ==True else 0)\n",
    "\n",
    "user_info_characteristics.index = data.keys()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Creating  __user_temp__ DataFrame just to store user features to be used for analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:35.197102Z",
     "start_time": "2018-10-23T21:40:35.152682Z"
    }
   },
   "outputs": [],
   "source": [
    "user_info_characteristics.head(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:36.647060Z",
     "start_time": "2018-10-23T21:40:35.881073Z"
    }
   },
   "outputs": [],
   "source": [
    "user_temp = user_info_characteristics[['verified','protected','description','contributors_enabled','geo_enabled','has_extended_profile','default_profile','favourites_count','followers_count','statuses_count','friends_count','listed_count','num_tweets','withheld_in_countries','date']]\n",
    "user_temp['followers_count/friends_count']= user_temp['followers_count']/user_temp['friends_count']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* __Lexical_Analysis : EMPATH__ and __Sentiment_Analysis:VADER__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:37.753849Z",
     "start_time": "2018-10-23T21:40:37.749749Z"
    }
   },
   "outputs": [],
   "source": [
    "# user_temp['sentiment_score'] = user_temp['description'].apply(lambda x:sentiment_analyzer.polarity_scores(x))\n",
    "# user_temp['lexicons']= user_temp['description'].apply(lambda x:lexicon.analyze(x,normalize = True))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:39.004978Z",
     "start_time": "2018-10-23T21:40:38.475542Z"
    }
   },
   "outputs": [],
   "source": [
    "from datetime import datetime\n",
    "def get_days(date):\n",
    "    date = date[0]\n",
    "    d  = re.findall(r'\\d+',date)\n",
    "    l  = '/'.join([d[1],d[2],d[0]])\n",
    "\n",
    "    date_format = \"%m/%d/%Y\"\n",
    "    a = datetime.strptime(l, date_format)\n",
    "    b = datetime.strptime('08/16/2018', date_format)\n",
    "    delta = b - a\n",
    "    return delta.days\n",
    "#def get_numTweets(x):\n",
    "user_temp['days']  = user_temp['date'].apply(get_days)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:39.554224Z",
     "start_time": "2018-10-23T21:40:39.530208Z"
    }
   },
   "outputs": [],
   "source": [
    "user_temp.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Normalization w.r.t to number of tweets per person:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:42.838306Z",
     "start_time": "2018-10-23T21:40:40.758145Z"
    }
   },
   "outputs": [],
   "source": [
    "user_temp['favourites_count/tweet']= user_temp['favourites_count']/user_temp['num_tweets']\n",
    "user_temp['followers_count/tweet']= user_temp['followers_count']/user_temp['num_tweets']\n",
    "user_temp['statuses_count/tweet']= user_temp['statuses_count']/user_temp['num_tweets']\n",
    "user_temp['friends_count/tweet']= user_temp['friends_count']/user_temp['num_tweets']\n",
    "user_temp['listed_count/tweet']= user_temp['listed_count']/user_temp['num_tweets']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:42.844437Z",
     "start_time": "2018-10-23T21:40:42.840239Z"
    }
   },
   "outputs": [],
   "source": [
    "# new_usertemp= user_temp[user_temp['followers_count/friends_count']!= np.inf]\n",
    "# print('Hate----Counter')\n",
    "# print(ttest_ind(new_usertemp.loc[set(lexicon_counter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'],new_usertemp.loc[set(lexicon_hate_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count']))\n",
    "# print(mannwhitneyu(new_usertemp.loc[set(lexicon_counter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'],new_usertemp.loc[set(lexicon_hate_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count']))\n",
    "# print()\n",
    "# print('Counter----NonCounter')\n",
    "# print(ttest_ind(new_usertemp.loc[set(lexicon_counter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'],new_usertemp.loc[set(lexicon_Noncounter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count']))\n",
    "# print(mannwhitneyu(new_usertemp.loc[set(lexicon_counter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'],new_usertemp.loc[set(lexicon_Noncounter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count']))\n",
    "# print()\n",
    "# print('Hate----NonCounter')\n",
    "# print(ttest_ind(new_usertemp.loc[set(lexicon_hate_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'],new_usertemp.loc[set(lexicon_Noncounter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count']))\n",
    "# print(mannwhitneyu(new_usertemp.loc[set(lexicon_hate_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'],new_usertemp.loc[set(lexicon_Noncounter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count']))\n",
    "# print()\n",
    "# Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(set(lexicon_Noncounter_user.keys()).intersection(new_usertemp.index)))],new_usertemp.loc[set(lexicon_Noncounter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'].tolist()],columns = ['Users','followers_count/friends_count'])\n",
    "# counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(set(lexicon_counter_user.keys()).intersection(new_usertemp.index)))],new_usertemp.loc[set(lexicon_counter_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'].tolist()],columns = ['Users','followers_count/friends_count'])\n",
    "# hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(set(lexicon_hate_user.keys()).intersection(new_usertemp.index)))],new_usertemp.loc[set(lexicon_hate_user.keys()).intersection(new_usertemp.index)]['followers_count/friends_count'].tolist()],columns = ['Users','followers_count/friends_count'])\n",
    "# Comb_df = pd.concat([Noncounter_df,counter_df,hate_df],ignore_index = True)\n",
    "# Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "# #             palette = {'Counter':'midnightblue','Hate':'crimson','NonCounter':'forestgreen'}\n",
    "# ax = sns.boxplot(x = 'Users',y ='followers_count/friends_count',data = Comb_df,width =0.3 ,linewidth = 2,showfliers = False)\n",
    "# plt.title('followers_count/friends_count')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Normalization w.r.t to number of days:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:45.009058Z",
     "start_time": "2018-10-23T21:40:42.846077Z"
    }
   },
   "outputs": [],
   "source": [
    "user_temp['favourites_count/day']= user_temp['favourites_count']/user_temp['days']\n",
    "user_temp['followers_count/day']= user_temp['followers_count']/user_temp['days']\n",
    "user_temp['statuses_count/day']= user_temp['statuses_count']/user_temp['days']\n",
    "user_temp['friends_count/day']= user_temp['friends_count']/user_temp['days']\n",
    "user_temp['listed_count/day']= user_temp['listed_count']/user_temp['days']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Checking out each row features by __loc feature in Pandas__:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-19T14:26:26.292255Z",
     "start_time": "2018-08-19T14:26:26.280645Z"
    }
   },
   "source": [
    "### Description Based Analysis:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:45.029374Z",
     "start_time": "2018-10-23T21:40:45.010695Z"
    }
   },
   "outputs": [],
   "source": [
    "#user_temp.to_json('Refined_Project_Dataset/user_temp.json',orient = 'records')\n",
    "# user_temp = pd.read_json('Refined_Project_Dataset/user_temp.json')\n",
    "# user_temp.index = data.keys()\n",
    "# user_temp = user_temp.apply(pd.to_numeric,errors = 'ignore')\n",
    "user_temp.head(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:45.176814Z",
     "start_time": "2018-10-23T21:40:45.032280Z"
    }
   },
   "outputs": [],
   "source": [
    "# num_date = []\n",
    "\n",
    "# for date in user_temp['date'][lexicon_hate_user.keys()].values.tolist():\n",
    "\n",
    "#     dte = date[0]\n",
    "#     d  = re.findall(r'\\d+',dte)\n",
    "#     l  = '/'.join([d[1],d[2],d[0]])\n",
    "\n",
    "#     date_format = \"%m/%d/%Y\"\n",
    "#     a = datetime.strptime(l, date_format)\n",
    "#     b = datetime.strptime('12/18/2017', date_format)\n",
    "#     if a>b:\n",
    "#         num_date.append(date)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:45.267077Z",
     "start_time": "2018-10-23T21:40:45.180084Z"
    }
   },
   "outputs": [],
   "source": [
    "# num_date = []\n",
    "# creation_dates = []\n",
    "# for i in tweets['reply_date_time'].values.tolist():\n",
    "#     date = re.findall('\\w+',i)\n",
    "# #     print(date)\n",
    "#     for i,j in enumerate(['Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec']):\n",
    "#         if len(date)!=2:  \n",
    "#             if date[1]==j:\n",
    "#                 date[1]=str(i+1)\n",
    "#         else:\n",
    "#             if date[0]==j:\n",
    "#                 date[0]=str(i+1)\n",
    "                \n",
    "#     try:\n",
    "#         if int(date[1]) in list(range(1,13)) and len(date)>2:\n",
    "#             date = '/'.join([date[1],date[0],date[2]])\n",
    "#             creation_dates.append(date)\n",
    "#         elif int(date[0]) in list(range(1,13)):\n",
    "#             date = '/'.join([date[0],date[1]])+'/2018'\n",
    "#             creation_dates.append(date)\n",
    "#     except:\n",
    "#         print()\n",
    "        \n",
    "# for dte in set(creation_dates):\n",
    "#     date_format = \"%m/%d/%Y\"\n",
    "#     a = datetime.strptime(dte, date_format)\n",
    "#     b = datetime.strptime('12/18/2017', date_format)\n",
    "#     if a>b:\n",
    "#         num_date.append(dte)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Description Based Analysis:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:45.392075Z",
     "start_time": "2018-10-23T21:40:45.270442Z"
    }
   },
   "outputs": [],
   "source": [
    "def get_hashtag_words(l):\n",
    "    all_descr = (' '.join(l)).lower()\n",
    "#     return Counter(re.findall(r'\\S+',all_descr))\n",
    "    return Counter(all_descr.split())\n",
    "hate_Description_hashtag_words = sorted(get_hashtag_words(user_temp.loc[sentiment_hate_user.keys()]['description'].tolist()).items(),key = lambda x: x[1],reverse = True)\n",
    "counter_Description_hashtag_words = sorted(get_hashtag_words(user_temp.loc[sentiment_counter_user.keys()]['description'].tolist()).items(),key = lambda x: x[1],reverse = True)\n",
    "Noncounter_Description_hashtag_words = sorted(get_hashtag_words(user_temp.loc[sentiment_Noncounter_user.keys()]['description'].tolist()).items(),key = lambda x: x[1],reverse = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# followers/day ......... Plotting "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:45.863962Z",
     "start_time": "2018-10-23T21:40:45.395008Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/navish/anaconda3/lib/python3.6/site-packages/numpy/core/_methods.py:112: RuntimeWarning: invalid value encountered in subtract\n",
      "  x = asanyarray(arr - arrmean)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                          favourites_count/day\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=292400.0, pvalue=0.43199909903588235)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=92523.5, pvalue=0.0005694057343323398)\n",
      "\n",
      "Hate----NonCounter\n",
      "Ttest_indResult(statistic=-1.1328540138115206, pvalue=0.2576662243156433)\n",
      "MannwhitneyuResult(statistic=40256.0, pvalue=0.000647903461469594)\n",
      "\n",
      "                          statuses_count/day\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=231699.0, pvalue=2.7464500674927732e-12)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=102853.0, pvalue=0.12319119324751043)\n",
      "\n",
      "Hate----NonCounter\n",
      "Ttest_indResult(statistic=2.8119553995694146, pvalue=0.005062985331558433)\n",
      "MannwhitneyuResult(statistic=39798.0, pvalue=0.0003235847977263939)\n",
      "\n",
      "                          followers_count/day\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=209557.5, pvalue=4.648272115477045e-21)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=104636.5, pvalue=0.21257271032010994)\n",
      "\n",
      "Hate----NonCounter\n",
      "Ttest_indResult(statistic=1.4334471539746259, pvalue=0.1521797925686616)\n",
      "MannwhitneyuResult(statistic=32078.0, pvalue=1.1684733966114967e-11)\n",
      "\n",
      "                          listed_count/day\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=232257.5, pvalue=3.0664973898867457e-12)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=100860.5, pvalue=0.05718146238644016)\n",
      "\n",
      "Hate----NonCounter\n",
      "Ttest_indResult(statistic=1.8916840007051388, pvalue=0.05894858732608934)\n",
      "MannwhitneyuResult(statistic=41038.0, pvalue=0.0019066169078028896)\n",
      "\n",
      "                          friends_count/day\n",
      "Hate----Counter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=268586.0, pvalue=0.002492365855382421)\n",
      "\n",
      "Counter----NonCounter\n",
      "Ttest_indResult(statistic=nan, pvalue=nan)\n",
      "MannwhitneyuResult(statistic=106526.0, pvalue=0.3392989862602931)\n",
      "\n",
      "Hate----NonCounter\n",
      "Ttest_indResult(statistic=0.6158585135152489, pvalue=0.5381895371797409)\n",
      "MannwhitneyuResult(statistic=44855.0, pvalue=0.1024773635382964)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for j in ['favourites_count/day','statuses_count/day','followers_count/day','listed_count/day','friends_count/day']:\n",
    "    _,pvalue1 = ttest_ind(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue2 = mannwhitneyu(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue3 = ttest_ind(user_temp.loc[lexicon_Noncounter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue4 = mannwhitneyu(user_temp.loc[lexicon_Noncounter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue5 = ttest_ind(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j])\n",
    "    _,pvalue6 = mannwhitneyu(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j])\n",
    "    if  pvalue2 < 0.01 or pvalue4 < 0.01 or pvalue6 < 0.01:\n",
    "            print(\"                         \",j)\n",
    "            print('Hate----Counter')\n",
    "            print(ttest_ind(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j]))\n",
    "            print(mannwhitneyu(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j]))\n",
    "            print()\n",
    "            print('Counter----NonCounter')\n",
    "            print(ttest_ind(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "            print(mannwhitneyu(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "            print()\n",
    "            print('Hate----NonCounter')\n",
    "            print(ttest_ind(user_temp.loc[lexicon_hate_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "            print(mannwhitneyu(user_temp.loc[lexicon_hate_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "            print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T22:36:33.616454Z",
     "start_time": "2018-10-23T22:36:32.069863Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/navish/anaconda3/lib/python3.6/site-packages/numpy/core/_methods.py:112: RuntimeWarning: invalid value encountered in subtract\n",
      "  x = asanyarray(arr - arrmean)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 612x144 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f,axes = plt.subplots(1,5,figsize=(8.5,2))\n",
    "col=0\n",
    "for j in ['favourites_count/day','statuses_count/day','followers_count/day','listed_count/day','friends_count/day']:\n",
    "    _,pvalue1 = ttest_ind(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue2 = mannwhitneyu(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue3 = ttest_ind(user_temp.loc[lexicon_Noncounter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue4 = mannwhitneyu(user_temp.loc[lexicon_Noncounter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j])\n",
    "    _,pvalue5 = ttest_ind(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j])\n",
    "    _,pvalue6 = mannwhitneyu(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j])\n",
    "    if  pvalue2 < 0.01 or pvalue4 < 0.01 or pvalue6 < 0.01:\n",
    "#             print(\"                         \",j)\n",
    "#             print('Hate----Counter')\n",
    "#             print(ttest_ind(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j]))\n",
    "#             print(mannwhitneyu(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_hate_user.keys()][j]))\n",
    "#             print()\n",
    "#             print('Counter----NonCounter')\n",
    "#             print(ttest_ind(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "#             print(mannwhitneyu(user_temp.loc[lexicon_counter_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "#             print()\n",
    "#             print('Hate----NonCounter')\n",
    "#             print(ttest_ind(user_temp.loc[lexicon_hate_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "#             print(mannwhitneyu(user_temp.loc[lexicon_hate_user.keys()][j],user_temp.loc[lexicon_Noncounter_user.keys()][j]))\n",
    "#             print()\n",
    "            Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(lexicon_Noncounter_user.keys()))],user_temp.loc[lexicon_Noncounter_user.keys()][j].tolist()],columns = ['Users',j])\n",
    "            counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(lexicon_counter_user.keys()))],user_temp.loc[lexicon_counter_user.keys()][j].tolist()],columns = ['Users',j])\n",
    "            hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(lexicon_hate_user.keys()))],user_temp.loc[lexicon_hate_user.keys()][j].tolist()],columns = ['Users',j])\n",
    "            Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "            Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "#             sns.boxplot(x = 'Users',y = j,data = Comb_df,width =0.5 ,linewidth = 1,showfliers = False,ax =axes[col])\n",
    "            sns.barplot(x = 'Users',y = j,data = Comb_df,errwidth=1.2,ax =axes[col],estimator = np.median,palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "            axes[col].set_xticklabels('')\n",
    "            axes[col].set_xlabel('')\n",
    "            axes[col].set_ylabel('')\n",
    "            axes[col].set_title('#'+re.findall(r'[a-zA-Z]+',j)[0]+'/'+re.findall(r'[a-zA-Z]+',j)[2])\n",
    "            col+=1\n",
    "            \n",
    "import matplotlib.patches as mpatches\n",
    "red_patch = mpatches.Patch(color='#c44e52', label='Hate')\n",
    "blue_patch = mpatches.Patch(color='#4c72b0', label='NonCounter')\n",
    "green_patch = mpatches.Patch(color='#55a868', label='Counter')\n",
    "f.legend(handles=[green_patch,red_patch],loc=9,ncol=3, frameon=False, borderaxespad =-0.25)\n",
    "f.subplots_adjust(top=2)\n",
    "f.tight_layout()\n",
    "f.savefig('Images/UserHistory_Barplots.eps', format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:47.757581Z",
     "start_time": "2018-10-23T21:40:47.747208Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(18, 7, 0)"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(user_temp.loc[lexicon_hate_user.keys()]['verified']==1).sum() , (user_temp.loc[lexicon_counter_user.keys()]['verified']==1).sum(),(user_temp.loc[lexicon_Noncounter_user.keys()]['verified']==1).sum() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:47.884076Z",
     "start_time": "2018-10-23T21:40:47.759198Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# ############################ According to certain Hues #########################\n",
    "# sns.set()\n",
    "# counter_df = pd.DataFrame(np.c_[np.array(['Counter' for i in range(len(user_temp.loc[lexicon_counter_user.keys()]['followers_count'].values))]),user_temp.loc[lexicon_counter_user.keys()][['followers_count','favourites_count','statuses_count','friends_count','listed_count','protected','verified']].values],columns = ['category','followers_count','favourites_count','statuses_count','friends_count','listed_count','protected','verified'])\n",
    "# hate_df = pd.DataFrame(np.c_[np.array(['Hate' for i in range(len(user_temp.loc[lexicon_hate_user.keys()]['followers_count'].values))]),user_temp.loc[lexicon_hate_user.keys()][['followers_count','favourites_count','statuses_count','friends_count','listed_count','protected','verified']].values],columns = ['category','followers_count','favourites_count','statuses_count','friends_count','listed_count','protected','verified'])\n",
    "# non_counter_df = pd.DataFrame(np.c_[np.array(['Non-Counter' for i in range(len(user_temp.loc[lexicon_Noncounter_user.keys()]['followers_count'].values))]),user_temp.loc[lexicon_Noncounter_user.keys()][['followers_count','favourites_count','statuses_count','friends_count','listed_count','protected','verified']].values],columns = ['category','followers_count','favourites_count','statuses_count','friends_count','listed_count','protected','verified'])\n",
    "# df = pd.concat([counter_df,hate_df,non_counter_df],ignore_index = True)\n",
    "# df = df.apply(pd.to_numeric, errors='ignore')\n",
    "# plt.figure(figsize=(3,3.2))\n",
    "# ax = sns.boxplot(x = 'category',y ='followers_count',hue= 'verified',data =df ,showfliers=False, linewidth=2,width = 0.5)\n",
    "# plt.show()\n",
    "\n",
    "# plt.figure(figsize=(3,3.2))\n",
    "# ax = sns.boxplot(x = 'category',y ='favourites_count',hue= 'verified',data =df ,showfliers=False, linewidth=2,width = 0.5)\n",
    "# plt.show()\n",
    "\n",
    "# plt.figure(figsize=(3,3.2))\n",
    "# ax = sns.boxplot(x = 'category',y ='statuses_count',hue= 'verified',data =df ,showfliers=False, linewidth=2,width = 0.5)\n",
    "# plt.show()\n",
    "\n",
    "# plt.figure(figsize=(3,3.2))\n",
    "# ax = sns.boxplot(x = 'category',y ='friends_count',hue= 'verified',data =df ,showfliers=False, linewidth=2,width = 0.5)\n",
    "# plt.show()\n",
    "\n",
    "# plt.figure(figsize=(3,3.2))\n",
    "# ax = sns.boxplot(x = 'category',y ='listed_count',hue= 'verified',data =df ,showfliers=False, linewidth=2,width = 0.5)\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:48.059642Z",
     "start_time": "2018-10-23T21:40:47.886992Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1827"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_temp['verified'].values.tolist().count(False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Date Based Analysis :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:48.343902Z",
     "start_time": "2018-10-23T21:40:48.062268Z"
    }
   },
   "outputs": [],
   "source": [
    "from datetime import datetime\n",
    "from dateutil import relativedelta\n",
    "from datetime import datetime\n",
    "all_dates = [j for j in list([i[0] for i in user_temp['date'].values.tolist()])]\n",
    "hate_dates = [j for j in list([i[0] for i in user_temp.loc[lexicon_hate_user,'date'].values.tolist()])]\n",
    "counter_dates =[j for j in list([i[0] for i in user_temp.loc[lexicon_counter_user,'date'].values.tolist()])]\n",
    "Non_counter_dates =[j for j in list([i[0] for i in user_temp.loc[lexicon_Noncounter_user,'date'].values.tolist()])]\n",
    "# from datetime import datetime\n",
    "# counter_dates = [datetime(int(re.findall('\\w+',i)[0]),int(re.findall('\\w+',i)[1]),int(re.findall('\\w+',i)[2])) for i in counter_dates] \n",
    "# hate_dates = [datetime(int(re.findall('\\w+',i)[0]),int(re.findall('\\w+',i)[1]),int(re.findall('\\w+',i)[2])) for i in hate_dates]\n",
    "def get_time_dict(l):\n",
    "    months = [] \n",
    "#     for i in range(130):\n",
    "#         months[i] = 0\n",
    "    for date in l:\n",
    "        date1 = datetime.strptime(date, '%Y-%m-%d')\n",
    "        date2 = datetime.strptime('2018-09-01', '%Y-%m-%d')\n",
    "        r = relativedelta.relativedelta(date2, date1)\n",
    "        months.append(r.months + (r.years+1)*12)\n",
    "#         if r.months + (r.years+1)*12 not in  months.keys():\n",
    "#             months[r.months + (r.years+1)*12] = 1\n",
    "#         else:\n",
    "#             months[r.months + (r.years+1)*12] += 1\n",
    "    return months\n",
    "hate_dates_list = get_time_dict(hate_dates)\n",
    "counter_dates_list = get_time_dict(counter_dates)\n",
    "Noncounter_dates_list = get_time_dict(Non_counter_dates)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:48.371465Z",
     "start_time": "2018-10-23T21:40:48.346040Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "450\n",
      "\n",
      "1039\n",
      "\n",
      "171\n"
     ]
    }
   ],
   "source": [
    "############################## Hate Dates before or in 2007#####################################\n",
    "h2,c2,nc2 = [],[],[]\n",
    "for i in hate_dates:\n",
    "    if int(re.findall(r'\\w+',i)[0])<=2017 and int(re.findall(r'\\w+',i)[1])<12:\n",
    "        h2.append(i)\n",
    "print(len(h2))\n",
    "print()\n",
    "for i in counter_dates:\n",
    "    if int(re.findall(r'\\w+',i)[0])<=2017 and int(re.findall(r'\\w+',i)[1])<12:\n",
    "        c2.append(i)\n",
    "print(len(c2))\n",
    "print()\n",
    "for i in Non_counter_dates:\n",
    "    if int(re.findall(r'\\w+',i)[0])<=2017 and int(re.findall(r'\\w+',i)[1])<12:\n",
    "        nc2.append(i)\n",
    "print(len(nc2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:48.524430Z",
     "start_time": "2018-10-23T21:40:48.373336Z"
    }
   },
   "outputs": [],
   "source": [
    "# hate_dates_Binny = []\n",
    "# counter_dates_Binny = []\n",
    "# for i in hate_dates:\n",
    "#     hate_dates_Binny.append(int(re.findall(r'\\w+',i)[0]))\n",
    "# for i in counter_dates:\n",
    "#     counter_dates_Binny.append(int(re.findall(r'\\w+',i)[0]))\n",
    "# np.savetxt('hate_dates_Binny.txt',hate_dates_Binny,fmt = '%1.0f')\n",
    "# np.savetxt('counter_dates_Binny.txt',counter_dates_Binny,fmt = '%1.0f')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Statistical tests :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:48.667296Z",
     "start_time": "2018-10-23T21:40:48.527358Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hate----Counter\n",
      "MannwhitneyuResult(statistic=270246.0, pvalue=0.004340122482279647)\n",
      "\n",
      "Counter----NonCounter\n",
      "MannwhitneyuResult(statistic=95929.5, pvalue=0.005184571933454095)\n",
      "\n",
      "Hate----NonCounter\n",
      "MannwhitneyuResult(statistic=37801.5, pvalue=1.0331586401625391e-05)\n"
     ]
    }
   ],
   "source": [
    "print('Hate----Counter')\n",
    "print(mannwhitneyu(hate_dates_list,counter_dates_list),end='\\n\\n')\n",
    "print('Counter----NonCounter')\n",
    "print(mannwhitneyu(Noncounter_dates_list,counter_dates_list),end='\\n\\n')\n",
    "print('Hate----NonCounter')\n",
    "print(mannwhitneyu(Noncounter_dates_list,hate_dates_list))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T22:37:01.419585Z",
     "start_time": "2018-10-23T22:37:01.191030Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x180 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set()\n",
    "Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(Noncounter_dates_list))],Noncounter_dates_list],columns = ['Users','Months'])\n",
    "counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(counter_dates_list))],counter_dates_list],columns = ['Users','Months'])\n",
    "hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(hate_dates_list))],hate_dates_list],columns = ['Users','Months'])\n",
    "Comb_df = pd.concat([counter_df,hate_df],ignore_index = True)\n",
    "Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "plt.figure(figsize = (4,2.5))\n",
    "ax = sns.violinplot(y = 'Users',x ='Months',data = Comb_df,width =0.6 ,linewidth =1 ,showfliers = False,palette = {'Hate':'#c44e52' ,'Counter':'#55a868'})\n",
    "ax.set_yticklabels('')\n",
    "ax.set_xticklabels(['','Sept 2018','Aug 2016','July 2014','June 2012','May 2010','Apr 2008','Mar 2006'])\n",
    "plt.ylabel('')\n",
    "plt.xlabel('')\n",
    "plt.xticks(rotation = 45)\n",
    "plt.title('Account Creation Dates')\n",
    "plt.tight_layout()\n",
    "plt.savefig('Images/Creation_Dates.eps', format='eps')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:48.961351Z",
     "start_time": "2018-10-23T21:40:48.958227Z"
    }
   },
   "outputs": [],
   "source": [
    "# import datetime as dt\n",
    "# Noncounter_dates_list = [dt.datetime.strptime(d,'%Y-%m-%d').date() for d in Non_counter_dates]\n",
    "# counter_dates_list = [dt.datetime.strptime(d,'%Y-%m-%d').date() for d in counter_dates]\n",
    "# hate_dates_list = [dt.datetime.strptime(d,'%Y-%m-%d').date() for d in hate_dates]\n",
    "# sns.set()\n",
    "# Noncounter_df = pd.DataFrame(np.c_[['NonCounter' for i in range(len(Noncounter_dates_list))],Noncounter_dates_list],columns = ['Users','Dates'])\n",
    "# counter_df = pd.DataFrame(np.c_[['Counter' for i in range(len(counter_dates_list))],counter_dates_list],columns = ['Users','Dates'])\n",
    "# hate_df = pd.DataFrame(np.c_[['Hate' for i in range(len(hate_dates_list))],hate_dates_list],columns = ['Users','Dates'])\n",
    "# Comb_df = pd.concat([Noncounter_df,counter_df,hate_df],ignore_index = True)\n",
    "# Comb_df = Comb_df.apply(pd.to_numeric, errors='ignore')\n",
    "# plt.figure(figsize = (5,1.5))\n",
    "# ax = sns.violinplot(y = 'Users',x ='Dates',data = Comb_df,width =0.6 ,linewidth =1 ,showfliers = False)\n",
    "# ax.set_yticklabels('')\n",
    "# plt.ylabel('')\n",
    "# plt.title('Creation Months from 1-Sept-2018')\n",
    "# # plt.savefig('Images/'+ 'Creation_Dates' +'.eps', format='eps')\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:49.123072Z",
     "start_time": "2018-10-23T21:40:48.971933Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(80.60510805500982, 75.87359307359307, 69.97340425531915)"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(hate_dates_list),np.mean(counter_dates_list),np.mean(Noncounter_dates_list)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  Likes, retweets :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:51.956142Z",
     "start_time": "2018-10-23T21:40:51.869778Z"
    }
   },
   "outputs": [],
   "source": [
    "with open('Refined_Project_Dataset/Hate_tweets_likes_retweets.json','r') as fp:\n",
    "    hate_tweets_likes_retweets = json.load(fp)\n",
    "with open('Refined_Project_Dataset/Counter_replies_retweets_likes.json','r') as fp:\n",
    "    counter_tweets_likes_retweets = json.load(fp)\n",
    "with open('Refined_Project_Dataset/NonCounter_replies_retweets_likes.json','r') as fp:\n",
    "    Noncounter_tweets_likes_retweets = json.load(fp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:52.416445Z",
     "start_time": "2018-10-23T21:40:52.409878Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(377, 841, 140)"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(hate_tweets_likes_retweets),len(counter_tweets_likes_retweets),len(Noncounter_tweets_likes_retweets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:52.812682Z",
     "start_time": "2018-10-23T21:40:52.807127Z"
    }
   },
   "outputs": [],
   "source": [
    "hate_tweets_likes_retweets[5],counter_tweets_likes_retweets[10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:53.899132Z",
     "start_time": "2018-10-23T21:40:53.890643Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "##### No. of Verified Users :\n",
    "len(user_temp[user_temp['verified']==1].index.tolist())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## For Verified Users : "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:54.481427Z",
     "start_time": "2018-10-23T21:40:54.429306Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(18, 7)"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "###### For Hate Users\n",
    "hId,cId = [],[]\n",
    "for i in  hate_usernames.intersection(set(user_temp[user_temp['verified']==1].index.tolist())):\n",
    "    hId.append(df[df['original_tweetusername']==i]['original_tweetid'].tolist()[0])\n",
    "for i in  counter_usernames.intersection(set(user_temp[user_temp['verified']==1].index.tolist())):\n",
    "    cId.append(df[df['replier_user_name']==i]['reply_id'].tolist()[0])\n",
    "len(hId),len(cId)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Verified Users /For all Users:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:55.140767Z",
     "start_time": "2018-10-23T21:40:55.112765Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "All Hates:\n",
      " retweet_count 6723  replies_count 3596 likes_count 23046\n",
      " retweet_mean 17.83289124668435  replies_mean 9.538461538461538 likes_mean 61.12997347480106 \n",
      " retweet_median 0.0  replies_median 3.0 likes_medain 2.0\n",
      "\n",
      "All counter:\n",
      " retweet_count 569  replies_count 389 likes_count 9690\n",
      " retweet_mean 0.6765755053507729  replies_mean 0.4625445897740785 likes_mean 11.521997621878716 \n",
      " retweet_median 0.0  replies_median 0.0 likes_medain 0.0\n",
      "\n",
      "All Noncounter:\n",
      " retweet_count 20  replies_count 44 likes_count 281\n",
      " retweet_mean 0.14285714285714285  replies_mean 0.3142857142857143 likes_mean 2.007142857142857 \n",
      " retweet_median 0.0  replies_median 0.0 likes_medain 0.0\n"
     ]
    }
   ],
   "source": [
    "#### For Hate User:\n",
    "retweets_count,likes_count,replies_count = 0,0,0\n",
    "retweets_list,likes_list,replies_list = [],[],[]\n",
    "for each_dict in hate_tweets_likes_retweets:\n",
    "        retweets_count+= int(each_dict['hate_no_retweets'])\n",
    "        retweets_list.append(int(each_dict['hate_no_retweets']))\n",
    "        replies_count+= int(each_dict['hate_no_replies'])\n",
    "        replies_list.append(int(each_dict['hate_no_replies']))\n",
    "        likes_count+= int(each_dict['hate_no_likes'])\n",
    "        likes_list.append(int(each_dict['hate_no_likes']))\n",
    "print('All Hates:\\n retweet_count {}  replies_count {} likes_count {}\\n retweet_mean {}  replies_mean {} likes_mean {} \\n retweet_median {}  replies_median {} likes_medain {}\\n'.format(retweets_count,replies_count,likes_count,np.mean(retweets_list),np.mean(replies_list),np.mean(likes_list),np.median(retweets_list),np.median(replies_list),np.median(likes_list))) \n",
    "# print(len(likes_list))\n",
    "#### For Counter User:\n",
    "retweets_count,likes_count,replies_count = 0,0,0\n",
    "retweets_list,likes_list,replies_list = [],[],[]\n",
    "for each_dict in counter_tweets_likes_retweets:\n",
    "        retweets_count+= int(each_dict['counter_no_retweets'])\n",
    "        retweets_list.append(int(each_dict['counter_no_retweets']))\n",
    "        replies_count+= int(each_dict['counter_no_replies'])\n",
    "        replies_list.append(int(each_dict['counter_no_replies']))\n",
    "        likes_count+= int(each_dict['counter_no_likes'])\n",
    "        likes_list.append(int(each_dict['counter_no_likes']))\n",
    "print('All counter:\\n retweet_count {}  replies_count {} likes_count {}\\n retweet_mean {}  replies_mean {} likes_mean {} \\n retweet_median {}  replies_median {} likes_medain {}\\n'.format(retweets_count,replies_count,likes_count,np.mean(retweets_list),np.mean(replies_list),np.mean(likes_list),np.median(retweets_list),np.median(replies_list),np.median(likes_list)))\n",
    "# print(len(likes_list))\n",
    "#### For NonCounter User:\n",
    "retweets_count,likes_count,replies_count = 0,0,0\n",
    "retweets_list,likes_list,replies_list = [],[],[]\n",
    "for each_dict in Noncounter_tweets_likes_retweets:\n",
    "        retweets_count+= int(each_dict['counter_no_retweets'])\n",
    "        retweets_list.append(int(each_dict['counter_no_retweets']))\n",
    "        replies_count+= int(each_dict['counter_no_replies'])\n",
    "        replies_list.append(int(each_dict['counter_no_replies']))\n",
    "        likes_count+= int(each_dict['counter_no_likes'])\n",
    "        likes_list.append(int(each_dict['counter_no_likes']))\n",
    "print('All Noncounter:\\n retweet_count {}  replies_count {} likes_count {}\\n retweet_mean {}  replies_mean {} likes_mean {} \\n retweet_median {}  replies_median {} likes_medain {}'.format(retweets_count,replies_count,likes_count,np.mean(retweets_list),np.mean(replies_list),np.mean(likes_list),np.median(retweets_list),np.median(replies_list),np.median(likes_list)))\n",
    "# print(len(likes_list))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:55.744194Z",
     "start_time": "2018-10-23T21:40:55.738789Z"
    }
   },
   "outputs": [],
   "source": [
    "Noncounter_tweets_likes_retweets[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-23T21:40:56.446073Z",
     "start_time": "2018-10-23T21:40:56.429760Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pic Hates:\n",
      " retweet_count 1116 replies_count 1323 likes_count 4443\n",
      " retweet_mean 139.5  replies_mean 165.375 likes_mean 555.375 \n",
      " retweet_median 13.5  replies_median 5.5 likes_medain 22.5\n",
      "\n",
      "Pic counter:\n",
      " retweet_count 75 replies_count 25 likes_count 858\n",
      " retweet_mean 0.8064516129032258  replies_mean 0.26881720430107525 likes_mean 9.225806451612904 \n",
      " retweet_median 0.0  replies_median 0.0 likes_medain 2.0\n",
      "\n",
      "Pic Noncounter:\n",
      " retweet_count 3 replies_count 7 likes_count 64\n",
      " retweet_mean 0.0967741935483871  replies_mean 0.22580645161290322 likes_mean 2.064516129032258 \n",
      " retweet_median 0.0  replies_median 0.0 likes_medain 1.0\n"
     ]
    }
   ],
   "source": [
    "#### For Hate User:\n",
    "retweets_count,likes_count,replies_count = 0,0,0\n",
    "retweets_list,likes_list,replies_list = [],[],[]\n",
    "for each_dict in hate_tweets_likes_retweets:\n",
    "    if int(each_dict['hate_id']) in hate_ids_pic:\n",
    "        retweets_count+= int(each_dict['hate_no_retweets'])\n",
    "        retweets_list.append(int(each_dict['hate_no_retweets']))\n",
    "        replies_count+= int(each_dict['hate_no_replies'])\n",
    "        replies_list.append(int(each_dict['hate_no_replies']))\n",
    "        likes_count+= int(each_dict['hate_no_likes'])\n",
    "        likes_list.append(int(each_dict['hate_no_likes']))\n",
    "print('Pic Hates:\\n retweet_count {} replies_count {} likes_count {}\\n retweet_mean {}  replies_mean {} likes_mean {} \\n retweet_median {}  replies_median {} likes_medain {}\\n'.format(retweets_count,replies_count,likes_count,np.mean(retweets_list),np.mean(replies_list),np.mean(likes_list),np.median(retweets_list),np.median(replies_list),np.median(likes_list))) \n",
    "# print(len(likes_list))\n",
    "#### For Counter User:\n",
    "retweets_count,likes_count,replies_count = 0,0,0\n",
    "retweets_list,likes_list,replies_list = [],[],[]\n",
    "for each_dict in counter_tweets_likes_retweets:\n",
    "    if int(each_dict['counter_id']) in counter_ids_pic:\n",
    "        retweets_count+= int(each_dict['counter_no_retweets'])\n",
    "        retweets_list.append(int(each_dict['counter_no_retweets']))\n",
    "        replies_count+= int(each_dict['counter_no_replies'])\n",
    "        replies_list.append(int(each_dict['counter_no_replies']))\n",
    "        likes_count+= int(each_dict['counter_no_likes'])\n",
    "        likes_list.append(int(each_dict['counter_no_likes']))\n",
    "print('Pic counter:\\n retweet_count {} replies_count {} likes_count {}\\n retweet_mean {}  replies_mean {} likes_mean {} \\n retweet_median {}  replies_median {} likes_medain {}\\n'.format(retweets_count,replies_count,likes_count,np.mean(retweets_list),np.mean(replies_list),np.mean(likes_list),np.median(retweets_list),np.median(replies_list),np.median(likes_list)))\n",
    "# print(len(likes_list))\n",
    "#### For NonCounter User:\n",
    "retweets_count,likes_count,replies_count = 0,0,0\n",
    "retweets_list,likes_list,replies_list = [],[],[]\n",
    "for each_dict in Noncounter_tweets_likes_retweets:\n",
    "    if int(each_dict['counter_id']) in NonCounter_ids_pic:\n",
    "        retweets_count+= int(each_dict['counter_no_retweets'])\n",
    "        retweets_list.append(int(each_dict['counter_no_retweets']))\n",
    "        replies_count+= int(each_dict['counter_no_replies'])\n",
    "        replies_list.append(int(each_dict['counter_no_replies']))\n",
    "        likes_count+= int(each_dict['counter_no_likes'])\n",
    "        likes_list.append(int(each_dict['counter_no_likes']))\n",
    "print('Pic Noncounter:\\n retweet_count {} replies_count {} likes_count {}\\n retweet_mean {}  replies_mean {} likes_mean {} \\n retweet_median {}  replies_median {} likes_medain {}'.format(retweets_count,replies_count,likes_count,np.mean(retweets_list),np.mean(replies_list),np.mean(likes_list),np.median(retweets_list),np.median(replies_list),np.median(likes_list)))\n",
    "# print(len(likes_list))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Retweet/Replies/Likes count per category [For all Users]:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-10-04T21:08:02.467988Z",
     "start_time": "2018-10-04T21:07:54.413621Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Category0\n",
      "Num_Tweets:  140\n",
      "retweets_count: 20 likes_count: 281 replies_count: 44\n",
      "retweets_count_avg: 0.14285714285714285 likes_count_avg: 2.007142857142857 replies_count_avg: 0.3142857142857143\n",
      "retweets_median: 0.0 likes_median: 0.0 replies_median: 0.0\n",
      "\n",
      "Category1\n",
      "Num_Tweets:  38\n",
      "retweets_count: 3 likes_count: 76 replies_count: 15\n",
      "retweets_count_avg: 0.07894736842105263 likes_count_avg: 2.0 replies_count_avg: 0.39473684210526316\n",
      "retweets_median: 0.0 likes_median: 0.5 replies_median: 0.0\n",
      "\n",
      "Category2\n",
      "Num_Tweets:  61\n",
      "retweets_count: 11 likes_count: 387 replies_count: 40\n",
      "retweets_count_avg: 0.18032786885245902 likes_count_avg: 6.344262295081967 replies_count_avg: 0.6557377049180327\n",
      "retweets_median: 0.0 likes_median: 0.0 replies_median: 0.0\n",
      "\n",
      "Category3\n",
      "Num_Tweets:  22\n",
      "retweets_count: 346 likes_count: 7652 replies_count: 20\n",
      "retweets_count_avg: 15.727272727272727 likes_count_avg: 347.8181818181818 replies_count_avg: 0.9090909090909091\n",
      "retweets_median: 0.0 likes_median: 1.0 replies_median: 0.0\n",
      "\n",
      "Category4\n",
      "Num_Tweets:  21\n",
      "retweets_count: 2 likes_count: 54 replies_count: 9\n",
      "retweets_count_avg: 0.09523809523809523 likes_count_avg: 2.5714285714285716 replies_count_avg: 0.42857142857142855\n",
      "retweets_median: 0.0 likes_median: 0.0 replies_median: 0.0\n",
      "\n",
      "Category5\n",
      "Num_Tweets:  45\n",
      "retweets_count: 4 likes_count: 25 replies_count: 19\n",
      "retweets_count_avg: 0.08888888888888889 likes_count_avg: 0.5555555555555556 replies_count_avg: 0.4222222222222222\n",
      "retweets_median: 0.0 likes_median: 0.0 replies_median: 0.0\n",
      "\n",
      "Category6\n",
      "Num_Tweets:  50\n",
      "retweets_count: 53 likes_count: 512 replies_count: 14\n",
      "retweets_count_avg: 1.06 likes_count_avg: 10.24 replies_count_avg: 0.28\n",
      "retweets_median: 0.0 likes_median: 2.5 replies_median: 0.0\n",
      "\n",
      "Category7\n",
      "Num_Tweets:  32\n",
      "retweets_count: 4 likes_count: 12 replies_count: 15\n",
      "retweets_count_avg: 0.125 likes_count_avg: 0.375 replies_count_avg: 0.46875\n",
      "retweets_median: 0.0 likes_median: 0.0 replies_median: 0.0\n",
      "\n",
      "Category8\n",
      "Num_Tweets:  93\n",
      "retweets_count: 18 likes_count: 152 replies_count: 46\n",
      "retweets_count_avg: 0.1935483870967742 likes_count_avg: 1.6344086021505377 replies_count_avg: 0.4946236559139785\n",
      "retweets_median: 0.0 likes_median: 0.0 replies_median: 0.0\n",
      "\n",
      "Category9\n",
      "Num_Tweets:  176\n",
      "retweets_count: 44 likes_count: 146 replies_count: 72\n",
      "retweets_count_avg: 0.25 likes_count_avg: 0.8295454545454546 replies_count_avg: 0.4090909090909091\n",
      "retweets_median: 0.0 likes_median: 0.0 replies_median: 0.0\n",
      "\n",
      "Category11\n",
      "Num_Tweets:  131\n",
      "retweets_count: 34 likes_count: 116 replies_count: 58\n",
      "retweets_count_avg: 0.2595419847328244 likes_count_avg: 0.8854961832061069 replies_count_avg: 0.44274809160305345\n",
      "retweets_median: 0.0 likes_median: 0.0 replies_median: 0.0\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for i in range(12):\n",
    "    retweets_count,likes_count,replies_count = 0,0,0\n",
    "    retweets_median,likes_median,replies_median = [],[],[]\n",
    "    if i==10:\n",
    "        continue\n",
    "    elif i==0:\n",
    "        for each_dict in Noncounter_tweets_likes_retweets:\n",
    "            retweets_count+=int(each_dict['counter_no_retweets'])\n",
    "            retweets_median.append(int(each_dict['counter_no_retweets']))\n",
    "            likes_count+=int(each_dict['counter_no_likes'])\n",
    "            likes_median.append(int(each_dict['counter_no_likes']))\n",
    "            replies_count+=int(each_dict['counter_no_replies'])\n",
    "            replies_median.append(int(each_dict['counter_no_replies']))\n",
    "        count = len(Noncounter_tweets_likes_retweets)\n",
    "        print('Category{}'.format(i))\n",
    "        print('Num_Tweets: ',count)\n",
    "        print('retweets_count: {} likes_count: {} replies_count: {}'.format(retweets_count,likes_count,replies_count))\n",
    "        print('retweets_count_avg: {} likes_count_avg: {} replies_count_avg: {}'.format(retweets_count/count,likes_count/count,replies_count/count))  \n",
    "        print('retweets_median: {} likes_median: {} replies_median: {}'.format(np.median(retweets_median),np.median(likes_median),np.median(replies_median)))\n",
    "        print()\n",
    "    else:\n",
    "        count=0\n",
    "        for each_dict in counter_tweets_likes_retweets:\n",
    "            if each_dict['counter_id'] in df[df['tag']==str(i)]['reply_id'].astype(str).tolist():\n",
    "                retweets_count+=int(each_dict['counter_no_retweets'])\n",
    "                retweets_median.append(int(each_dict['counter_no_retweets']))\n",
    "                likes_count+=int(each_dict['counter_no_likes'])\n",
    "                likes_median.append(int(each_dict['counter_no_likes']))\n",
    "                replies_count+=int(each_dict['counter_no_replies'])\n",
    "                replies_median.append(int(each_dict['counter_no_replies']))\n",
    "                count += 1\n",
    "        print('Category{}'.format(i))\n",
    "        print('Num_Tweets: ',count)\n",
    "        print('retweets_count: {} likes_count: {} replies_count: {}'.format(retweets_count,likes_count,replies_count))\n",
    "        print('retweets_count_avg: {} likes_count_avg: {} replies_count_avg: {}'.format(retweets_count/count,likes_count/count,replies_count/count))                             \n",
    "        print('retweets_median: {} likes_median: {} replies_median: {}'.format(np.median(retweets_median),np.median(likes_median),np.median(replies_median)))\n",
    "        print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}