unit-project/3/notebooks/unit-project-3-starter-code.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# DS-SF-30 | Unit Project 3: Machine Learning Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this project, you will perform a logistic regression on the admissions data we've been working with in Unit Projects 1 and 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "pd.set_option('display.max_rows', 10)\n",
    "pd.set_option('display.max_columns', 10)\n",
    "pd.set_option('display.notebook_repr_html', True)\n",
    "\n",
    "import statsmodels.formula.api as smf\n",
    "\n",
    "from sklearn import linear_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv(os.path.join('..', '..', 'dataset', 'dataset-ucla-admissions.csv'))\n",
    "df.dropna(inplace = True)\n",
    "\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part A.  Frequency Table"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 1.  Create a frequency table for `prestige` and whether an applicant was admitted."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part B.  Variable Transformations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 2.  Create a one-hot encoding for `prestige`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 3.  How many of these binary variables do we need for modeling?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 4.  Why are we doing this?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 5.  Add all these binary variables in the dataset and remove the now redundant `prestige` feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part C.  Hand calculating odds ratios"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's develop our intuition about expected outcomes by hand calculating odds ratios."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 6.  Create a frequency table for `prestige = 1` and whether an applicant was admitted."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 7.  Use the frequency table above to calculate the odds of being admitted to graduate school for applicants that attended the most prestigious undergraduate schools."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 8.  Now calculate the odds of admission for undergraduates who did not attend a #1 ranked college."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 9.  Finally, what's the odds ratio?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 10.  Write this finding in a sentence."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 11.  Use the frequency table above to calculate the odds of being admitted to graduate school for applicants that attended the least prestigious undergraduate schools.  Then calculate their odds ratio of being admitted to UCLA.  Finally, write this finding in a sentence."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part C. Analysis using `statsmodels`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 12.  Fit a logistic regression model predicting admission into UCLA using `gre`, `gpa`, and the `prestige` of the undergraduate schools.  Use the highest prestige undergraduate schools as your reference point."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 13.  Print the model's summary results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 14.  What are the odds ratios of the different features and their 95% confidence intervals?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 15.  Interpret the odds ratio for `prestige = 2`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 16.  Interpret the odds ratio of `gpa`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 17.  Assuming a student with a GRE of 800 and a GPA of 4.  What is his/her probability of admission  if he/she come from a tier-1, tier-2, tier-3, or tier-4 undergraduate school?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part D. Moving the model from `statsmodels` to `sklearn`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 18.  Let's assume we are satisfied with our model.  Remodel it (same features) using `sklearn`.  When creating the logistic regression model with `LogisticRegression(C = 10 ** 2)`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 19.  What are the odds ratios for the different variables and how do they compare with the odds ratios calculated with `statsmodels`?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Question 20.  Again, assuming a student with a GRE of 800 and a GPA of 4.  What is his/her probability of admission  if he/she come from a tier-1, tier-2, tier-3, or tier-4 undergraduate school?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# TODO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer:"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}