Facing Issue with DASH Board in Streamlit (Titanic/Kashti Case study)

[AzamBukhari]

AoAlaikum

While working on the video tutorial of Data Science DASH Board in Streamlit (Titanic/Kashti Case study)
(code is pasted below)
[Video Link:] https://www.youtube.com/watch?v=hy0A6wxlBuM&list=PL9XvIvvVL50FG8TUsLScsLMmJDacK36x2&index=2

I am facing two problems:

1. I am unable to display metrices values in the right column of Streamlit
2. My model is unable to get fit., since the feature variable X is empty at the time of initialization. However it works with BABA-G video

I have taken help from chatGPT , but no hope ..

Can any one of you try the code and see if they are getting the same issue

---

import streamlit as st
import seaborn as sns
import pandas as pd
from sklearn.ensemble import RandomForestRegressor
from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score

##make containers
#just like webpage frames

header = st.container()
data_sets = st.container()
features = st.container()
model_training = st.container()

with header:
    st.title("Kashti ki app")
    st.text("In this project we will workon kashti data")
    
with data_sets:
    st.header("Kashti Doob Gui, haawww")
    st.text("We woill work with titanic data set")
    df=sns.load_dataset("titanic")
    st.write(df.head(10))
    df = df.dropna()
    st.write(df.head(10))
    
    st.subheader("Sambha, aray oo sambha kitney adami they?")
    st.bar_chart(df['sex'].value_counts())
    
    #other plot
    st.subheader("Class  K Hisaab b se Ferq")
    st.bar_chart(df['class'].value_counts())
    
    #bar plot    
    st.bar_chart(df['age'].sample(10))

  
with features:
    st.header("These are our app features")
    st.text("Ainwaee bauhat sarey features add kerte ahain, asaan hi hai")
    
    st.markdown("> 1. **Feature 1:** will tell us pata nahin kya")
    st.markdown("> 2. **Feature 2:** will tell us pata nahin kya")
    
    
with model_training:
    st.header("Kashti Walon ka kya buna>? Model training")
    st.text("In this project we will workon kashti data")
    
    ##making columns
    input, display = st.columns(2)

    #pehley column mai humarey selection point hun
    max_depth = input.slider("How many people Do you know?", min_value=10, max_value=100, value=20, step=5 )
    
    #n_estimators
    n_estimators = input.selectbox("How many trees shoukd be there in a rndom forest", options=[50,100, 200, 400, 'No Limit'])


    # adding list of features
    input.write(df.columns)
    
    #input features from users
    input_features = input.text_input("which feature we should use?")
    
    
    
    #machine learning model
    model = RandomForestRegressor(max_depth=max_depth, n_estimators=n_estimators)
    #yahan per ham aik condition lugain gai
    if n_estimators == 'No Limit':
        random_r = RandomForestRegressor(max_depth==max_depth)
    else:
        random_r = RandomForestRegressor(max_depth=max_depth, n_estimators=n_estimators)
        
    #define X & y, input & output
    X = df[[input_features]]
    y = df[['fare']]
    
    #fit our model
    model.fit(X, y)    
    pred = model.predict(y)
    
with display.right:
        # Second column
    
    #display model metrices
    display.subheader("Mean Absolute error of the model is: ")
    display.write(mean_absolute_error, y, pred)
    display.subheader("Mean Squared error of the model is: ")
    display.write(mean_squared_error, y, pred)
    display.subheader("r2 Score error of the model is: ")
    display.write(r2_score, y, pred)

[mahnoor-byte]

I am facing similar issue.

[Muhammad-Ali-Butt]

walaikum asalam azam bhai i had the same error but when i put fare in the input text thingy that somehow worked

[AzamBukhari]

Thanks Muammad Ali , I am more interested in the reason. Why BABA_g ka chul gaya aur huamin work around kerna perh raha hai

[Muhammad-Ali-Butt]

Sure, I'll try this one more time thanks for the reminder 👍

[Muhammad-Ali-Butt]

Thanks Muammad Ali , I am more interested in the reason. Why BABA_g ka chul gaya aur huamin work around kerna perh raha hai

@AzamBukhari @seharfatimaa i think we had to work around due to some changes in streamlit etc. you can confirm it from @AammarTufail also i figured an easy way to make it work with a simple if statement when defining X and y like this:

[Muhammad-Ali-Butt]

Explained with comments for anyone who is facing this problem

# define X and y
if input_features in df.columns:   # means agar input_features i.e jo user ne type kiye are in df.columns then:
    X = df[[input_features]]
    y = df[['fare']]

    # Fit model
    model.fit(X, y)
    pred = model.predict(X)

else:                # aur agar input_features are not in df.columns then write the error
    st.write(
        'The feature you entered is not present in the data. Please enter a valid feature name.') 

[AzamBukhari]

@Muhammad-Ali-Butt Thanks bhuiyya, still I am getting error.
If it is possible , please share your side of the code.

The other issue I am facing is I am not able to display any data in the 2nd column.

Must be doing or missing something.

[Muhammad-Ali-Butt]

sorry for delayed response azam bhai i was kinda busy and not checking github

import streamlit as st
import pandas as pd
import seaborn as sns
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
# ? Make containers
header = st.container()
data_sets = st.container()
featuress = st.container()
model_training = st.container()

with header:
    st.title('Kashti ki app')
    st.text('In this project we will work on kashti data')

with data_sets:
    st.header('Kashti doob gai')
    st.text('We will work with titanic data set')
    # ? Import data
    df = sns.load_dataset('titanic')
    df = df.dropna()  # filling missing values
    st.write(df.head(10))

    st.subheader('Barchart of sexes')
    st.bar_chart(df['sex'].value_counts())

    # ? Other plot
    st.subheader('Barchart of classes')
    st.bar_chart(df['class'].value_counts())

    # ? barplot
    st.bar_chart(df['age'].sample(10))

with featuress:
    st.header('These are our app Features:')
    st.text('We add features here')

    st.markdown('1. **Feature 1:** This will tell us some ting')
    st.markdown('2. **Feature 2:** This will tell us some ting')

with model_training:
    st.header('Kashti waalo ka kia bana')
    st.text('In this project we can train our model')
    # ? making columns
    inputing, display = st.columns(2)
    # ? In 1st column we add selection points
    max_depth = inputing.slider(
        'How many people do you know', min_value=10, max_value=100, value=20, step=5)

# ? n_estimators
n_estimators = inputing.selectbox('How many Trees should be in RF', options=[
    50, 100, 200, 300, 'NO LIMIT'])

# Input features from user
input_features = inputing.text_input('Which feature we should use')
# inputing.write(df.head(3))
# Machine learning model
model = RandomForestRegressor(max_depth=max_depth, n_estimators=n_estimators)
#! If user select NO LIMIT then
if n_estimators == 'NO LIMIT':
    model = RandomForestRegressor(max_depth=max_depth)
else:
    model = RandomForestRegressor(
        max_depth=max_depth, n_estimators=n_estimators)

# define X and y
X = df[[input_features]]
y = df['fare']

# st.write(X.shape, y.shape)

# # Fit model
model.fit(X, y)
pred = model.predict(X)
# else:
# st.write(
# 'The feature you entered is not present in the data. Please enter a valid feature name.')

# Display metrics
display.subheader('Mean absolute error of the model is:')
display.write(mean_absolute_error(y, pred))
display.subheader('Mean squared error of the model is:')
display.write(mean_squared_error(y, pred))
display.subheader('R2 squared of the model is:')
display.write(r2_score(y, pred))