DevanPutra/Churn_Prediction

1import streamlit as st2import pandas as pd3import numpy as np4import joblib5import tensorflow as tf6from tensorflow.keras.models import Sequential, Model, load_model7 8# Load model and feature names9with open('pipeline.pkl', 'rb') as file_1:10    pipeline= joblib.load(file_1)11 12model_func_tune= load_model('model.h5')13 14# Set up app title and header image15st.set_page_config(page_title='Churn Rate Prediction')16st.title('Churn Rate Prediction')17st.image('https://i.imgur.com/rToAL0W.png', use_column_width=True)18 19# Collect user input using sliders20st.subheader('Enter Customers Information:')21membership_category = st.selectbox('membership_category', tuple(['Silver Membership', 'Gold Membership', 'No Membership',22                                                                 'Basic Membership', 'Platinum Membership', 'Premium Membership']))23joined_through_referral = st.selectbox('joined_through_referral', tuple(['Yes', 'No']))24feedback = st.selectbox('feedback', tuple(['Poor Product Quality', 'Poor Website', 'No reason specified',25       'Quality Customer Care', 'Poor Customer Service',26       'Reasonable Price', 'User Friendly Website', 'Too many ads',27       'Products always in Stock']))28avg_transaction_value = st.slider ('avg_transaction_value $', 0, 100000)29avg_frequency_login_days = st.slider ('avg_frequency_login_days ', 0, 50)30points_in_wallet = st.slider ('points_in_wallet $', 0, 1500)31 32 33# Generate new DataFrame based on user input34new_data = pd.DataFrame({35    'membership_category': [membership_category],36    'joined_through_referral': [joined_through_referral],37    'feedback': [feedback],38    'avg_transaction_value': [avg_transaction_value],39    'avg_frequency_login_days': [avg_frequency_login_days],40    'points_in_wallet': [points_in_wallet]41})42df= pipeline.transform(new_data)43# Make prediction and display result44if st.button('Predict'):45    y_pred_func_tuned = model_func_tune.predict(df)46    y_pred_func_tuned = np.where(y_pred_func_tuned >=0.5, 1, 0)  47    if y_pred_func_tuned == 1:48        message = "From the Customers information, it seems that the customers is churn."49        color = 'red'50    else:51        message = "From the Customers information, it seems that the customers is not churn."52        color = 'green'53    st.subheader('Prediction:')54    st.write(message, unsafe_allow_html=True, )55    st.markdown(f'<h1 style="color:{color};text-align:center">{y_pred_func_tuned}</h1>', unsafe_allow_html=True)56