import pandas as pd from transformers import pipeline import streamlit as st import datetime from huggingface_hub import hf_hub_download import joblib REPO_ID = "AlbieCofie/predict-customer-churn" num_imputer = joblib.load( hf_hub_download(repo_id=REPO_ID, filename="numerical_imputer.joblib") ) cat_imputer = joblib.load( hf_hub_download(repo_id=REPO_ID, filename="categorical_imputer.joblib") ) encoder = joblib.load( hf_hub_download(repo_id=REPO_ID, filename="encoder.joblib") ) scaler = joblib.load( hf_hub_download(repo_id=REPO_ID, filename="scaler.joblib") ) model = joblib.load( hf_hub_download(repo_id=REPO_ID, filename="Final_model.joblib") ) # Create a function that applies the ML pipeline and makes predictions def predict(gender,SeniorCitizen,Partner,Dependents, tenure, PhoneService,MultipleLines, InternetService,OnlineSecurity,OnlineBackup,DeviceProtection,TechSupport,StreamingTV,StreamingMovies, Contract,PaperlessBilling,PaymentMethod,MonthlyCharges,TotalCharges): # Create a dataframe with the input data input_df = pd.DataFrame({ 'gender': [gender], 'SeniorCitizen': [SeniorCitizen], 'Partner': [Partner], 'Dependents': [Dependents], 'tenure': [tenure], 'PhoneService': [PhoneService], 'MultipleLines': [MultipleLines], 'InternetService': [InternetService], 'OnlineSecurity': [OnlineSecurity], 'OnlineBackup': [OnlineBackup], 'DeviceProtection': [DeviceProtection], 'TechSupport': [TechSupport], 'StreamingTV': [StreamingTV], 'StreamingMovies': [StreamingMovies], 'Contract': [Contract], 'PaperlessBilling': [PaperlessBilling], 'PaymentMethod': [PaymentMethod], 'MonthlyCharges': [MonthlyCharges], 'TotalCharges': [TotalCharges] }) # Selecting categorical and numerical columns separately cat_columns = [col for col in input_df.columns if input_df[col].dtype == 'object'] num_columns = [col for col in input_df.columns if input_df[col].dtype != 'object'] # Apply the imputers on the input data input_df_imputed_cat = cat_imputer.transform(input_df[cat_columns]) input_df_imputed_num = num_imputer.transform(input_df[num_columns]) # Encode the categorical columns input_encoded_df = pd.DataFrame(encoder.transform(input_df_imputed_cat).toarray(), columns=encoder.get_feature_names_out(cat_columns)) # Scale the numerical columns input_df_scaled = scaler.transform(input_df_imputed_num) input_scaled_df = pd.DataFrame(input_df_scaled , columns = num_columns) #joining the cat encoded and num scaled final_df = pd.concat([input_encoded_df, input_scaled_df], axis=1) final_df = final_df.reindex(columns=['SeniorCitizen','tenure','MonthlyCharges','TotalCharges', 'gender_Female','gender_Male','Partner_No','Partner_Yes','Dependents_No','Dependents_Yes','PhoneService_No', 'PhoneService_Yes','MultipleLines_No','MultipleLines_Yes','InternetService_DSL','InternetService_Fiber optic', 'InternetService_No','OnlineSecurity_No','OnlineSecurity_Yes','OnlineBackup_No','OnlineBackup_Yes','DeviceProtection_No', 'DeviceProtection_Yes','TechSupport_No','TechSupport_Yes','StreamingTV_No','StreamingTV_Yes','StreamingMovies_No', 'StreamingMovies_Yes','Contract_Month-to-month','Contract_One year','Contract_Two year','PaperlessBilling_No', 'PaperlessBilling_Yes','PaymentMethod_Bank transfer (automatic)','PaymentMethod_Credit card (automatic)','PaymentMethod_Electronic check', 'PaymentMethod_Mailed check']) # Make predictions using the model predictions = model.predict(final_df)[0] #prediction = model.predict(final_df)[0] # Make predictions using the model #predictions = model.predict(final_df) # Convert the numpy array to an integer #prediction_label = int(predictions.item()) prediction_label = "Beware!!! This customer is likely to Churn" if predictions.item() == "Yes" else "This customer is Not likely churn" return prediction_label #return predictions if 'clicked' not in st.session_state: st.session_state.clicked = False def click_button(): st.session_state.clicked = True st.title("CUSTOMER CHURN PREDICTION APP") with st.form(key="customer-information"): st.markdown("This app predicts whether a customer will leave your company or not. Enter the details of the customer below to see the result") gender = st.radio('Select your gender', ('male', 'female')) SeniorCitizen = st.radio("Are you a Seniorcitizen; No=0 and Yes=1", ('0', '1')) Partner = st.radio('Do you have Partner', ('Yes', 'No')) Dependents = st.selectbox('Do you have any Dependents?', ('No', 'Yes')) tenure = st.number_input('Lenght of tenure (no. of months with Telco)', min_value=0, max_value=90, value=1, step=1) PhoneService = st.radio('Do you have PhoneService? ', ('No', 'Yes')) MultipleLines = st.radio('Do you have MultipleLines', ('No', 'Yes')) InternetService = st.radio('Do you have InternetService', ('DSL', 'Fiber optic', 'No')) OnlineSecurity = st.radio('Do you have OnlineSecurity?', ('No', 'Yes')) OnlineBackup = st.radio('Do you have OnlineBackup?', ('No', 'Yes')) DeviceProtection = st.radio('Do you have DeviceProtection?', ('No', 'Yes')) TechSupport = st.radio('Do you have TechSupport?', ('No', 'Yes')) StreamingTV = st.radio('Do you have StreamingTV?', ('No', 'Yes')) StreamingMovies = st.radio('Do you have StreamingMovies?', ('No', 'Yes')) Contract = st.selectbox('which Contract do you use?', ('Month-to-month', 'One year', 'Two year')) PaperlessBilling = st.radio('Do you prefer PaperlessBilling?', ('Yes', 'No')) PaymentMethod = st.selectbox('Which PaymentMethod do you prefer?', ('Electronic check', 'Mailed check', 'Bank transfer (automatic)', 'Credit card (automatic)')) MonthlyCharges = st.number_input("Enter monthly charges (the range should between 0-120)") TotalCharges = st.number_input("Enter total charges (the range should between 0-10.000)") st.form_submit_button('Predict', on_click=click_button) if st.session_state.clicked: # The message and nested widget will remain on the page predict(gender,SeniorCitizen,Partner,Dependents, tenure, PhoneService,MultipleLines, InternetService,OnlineSecurity,OnlineBackup,DeviceProtection,TechSupport,StreamingTV,StreamingMovies, Contract,PaperlessBilling,PaymentMethod,MonthlyCharges,TotalCharges)