FKBaffour/Customer_Churn_Prediction_App

1# Importing required Libraries2from IPython.utils.py3compat import encode3import gradio as gr4import numpy as np5import pandas as pd6import pickle7 8 9# Loading Machine Learning Objects10def load_saved_objets(filepath='ML_items'):11    "Function to load saved objects"12 13    with open(filepath, 'rb') as file:14        loaded_object = pickle.load(file)15    16    return loaded_object17 18# Instantiating ML_items19loaded_object = load_saved_objets()20pipeline_of_my_app = loaded_object["pipeline"]21num_cols = loaded_object['numeric_columns']22cat_cols = loaded_object['categorical_columns']23encoder_categories = loaded_object["encoder_categories"]24 25# Main function to collect the inputs process them and outpuT the predicition26def predict_churn(27    TotalCharges,28    MonthlyCharges,29    tenure, 30    StreamingTV,31    PaperlessBilling,32    DeviceProtection,33    TechSupport,34    InternetService,35    OnlineSecurity,36    StreamingMovies,37    PaymentMethod,38    Dependents,39    Parter,40    tenure_group,41    OnlineBackup,42    gender,43    SeniorCitizen,44    MultipleLines,45    Contract,46    PhoneService,47):48    49    df = pd.DataFrame(50        [51            [52                TotalCharges,53                MonthlyCharges,54                tenure, 55                StreamingTV,56                PaperlessBilling,57                DeviceProtection,58                TechSupport,59                InternetService,60                OnlineSecurity,61                StreamingMovies,62                PaymentMethod,63                Dependents,64                Parter,65                tenure_group,66                OnlineBackup,67                gender,68                SeniorCitizen,69                MultipleLines,70                Contract,71                PhoneService,72            ]73        ],  74        columns= num_cols + cat_cols,75    ).replace("", np.nan)76    77    df[cat_cols] = df[cat_cols].astype("object")78    79    # Passing data to pipeline to make prediction80    output = pipeline_of_my_app.predict(df)81    82    # Labelling Model output83    if output == 0:84        model_output = "No"85    else:86        model_output = "Yes"87 88    return model_output89 90 91# Setting up app interface and data inputs92inputs = []93 94with gr.Blocks() as demo:95    96    # Setting Titles for App97    gr.Markdown("<h2 style='text-align: center;'> Customer Churn Prediction App </h2> ", unsafe_allow_html=True)98    gr.Markdown("<h6 style='text-align: center;'> (Fill in the details below and click on PREDICT button to make a prediction for Customer Churn) </h6> ", unsafe_allow_html=True)   99    100    with gr.Column(): #main frame 101        102        with gr.Row(): #col 1 : for num features103 104            for i in num_cols:105                inputs.append(gr.Number(label=f"Input {i} "))106        107        with gr.Row(): #col 2 : for cat features108 109            for (lab, choices) in zip(cat_cols, encoder_categories):110                inputs.append(gr.inputs.Dropdown(111                choices=choices.tolist(),112                type="value",113                label=f"Select {lab}",114                default=choices.tolist()[0],))115    # Setting up preediction Button116    with gr.Row():117        make_prediction = gr.Button("Predict")118    119    # Setting up prediction output Row120    with gr.Row():121        output_prediction = gr.Text(label="Will Customer Churn?")122    make_prediction.click(predict_churn, inputs, output_prediction)123 124# Launching app125demo.launch()