Update app.py

app.py

@@ -1,6 +1,7 @@
 import gradio as gr
 import pandas as pd
 import joblib
+import numpy as np
 from sklearn.pipeline import Pipeline
 from sklearn.impute import SimpleImputer
 from sklearn.compose import ColumnTransformer
@@ -39,6 +40,88 @@ def predict(gender, SeniorCitizen, Partner, Dependents, Contract, tenure, Monthl
     })
 
 
+        # Make predictions using the loaded logistic regression model
+        #predict probabilities
+    predictions = full_pipeline.predict_proba(input_data)
+    #take the index of the maximum probability
+    index=np.argmax(predictions)
+    higher_pred_prob=round((predictions[0][index])*100)
+
+
+    #return predictions[0]
+    print(f'[Info] Predicted probabilities{predictions},{full_pipeline.classes_}')
+    if full_pipeline.classes_[index] == "Yes":
+        return f"This Customer is likely to Churn\nWe are {higher_pred_prob}% confident about this prediction"
+    else:
+        return f"This Customer is Not likely to Churn \nWe are {higher_pred_prob}% confident about this prediction"
+    
+# Setting Gradio App Interface
+with gr.Blocks(css=".gradio-container {background-color: grey}",theme=gr.themes.Base(primary_hue='blue'),title='Uriel') as demo:
+    gr.Markdown("# Teleco Customer Churn Prediction #\n*This App allows the user to predict whether a customer will churn or not by entering values in the given fields. Any field left blank takes the default value.*")
+    
+    # Receiving ALL Input Data here
+    gr.Markdown("**Demographic Data**")
+    with gr.Row():
+        gender = gr.Dropdown(label="Gender", choices=["Male", "Female"])
+        SeniorCitizen = gr.Radio(label="Senior Citizen", choices=[1, 0])
+        Partner = gr.Radio(label="Partner", choices=["Yes", "No"])
+        Dependents = gr.Radio(label="Dependents", choices=["Yes", "No"])
+
+    gr.Markdown("**Service Length and Charges (USD)**")
+    with gr.Row():
+        Contract = gr.Dropdown(label="Contract", choices=["Month-to-month", "One year", "Two year"])
+        tenure = gr.Slider(label="Tenure (months)", minimum=1, step=1, interactive=True)
+        MonthlyCharges = gr.Slider(label="Monthly Charges", step=0.05)
+        TotalCharges = gr.Slider(label="Total Charges", step=0.05)
+
+    # Phone Service Usage part
+    gr.Markdown("**Phone Service Usage**")
+    with gr.Row():
+        PhoneService = gr.Radio(label="Phone Service", choices=["Yes", "No"])
+        MultipleLines = gr.Dropdown(label="Multiple Lines", choices=[
+                                    "Yes", "No", "No phone service"])
+
+    # Internet Service Usage part
+    gr.Markdown("**Internet Service Usage**")
+    with gr.Row():
+        InternetService = gr.Dropdown(label="Internet Service", choices=["DSL", "Fiber optic", "No"])
+        OnlineSecurity = gr.Dropdown(label="Online Security", choices=["Yes", "No", "No internet service"])
+        OnlineBackup = gr.Dropdown(label="Online Backup", choices=["Yes", "No", "No internet service"])
+        DeviceProtection = gr.Dropdown(label="Device Protection", choices=["Yes", "No", "No internet service"])
+        TechSupport = gr.Dropdown(label="Tech Support", choices=["Yes", "No", "No internet service"])
+        StreamingTV = gr.Dropdown(label="TV Streaming", choices=["Yes", "No", "No internet service"])
+        StreamingMovies = gr.Dropdown(label="Movie Streaming", choices=["Yes", "No", "No internet service"])
+
+    # Billing and Payment part
+    gr.Markdown("**Billing and Payment**")
+    with gr.Row():
+        PaperlessBilling = gr.Radio(
+            label="Paperless Billing", choices=["Yes", "No"])
+        PaymentMethod = gr.Dropdown(label="Payment Method", choices=["Electronic check", "Mailed check", "Bank transfer (automatic)", "Credit card (automatic)"])
+
+    # Output Prediction
+    output = gr.Text(label="Outcome")
+    submit_button = gr.Button("Predict")
+    
+    submit_button.click(fn= predict,
+                        outputs= output,
+                        inputs=[gender, SeniorCitizen, Partner, Dependents, Contract, tenure, MonthlyCharges, TotalCharges, PaymentMethod, PhoneService, MultipleLines, InternetService, OnlineSecurity, OnlineBackup, DeviceProtection, TechSupport, StreamingTV, StreamingMovies, PaperlessBilling],
+    
+    ),
+    
+    # Add the reset and flag buttons
+    
+    def clear():
+        output.value = ""
+        return 'Predicted values have been reset'
+         
+    clear_btn = gr.Button("Reset", variant="primary")
+    clear_btn.click(fn=clear, inputs=None, outputs=output)
+        
+ 
+demo.launch(inbrowser = True)
+
+
         # Make predictions using the loaded logistic regression model
     predictions = full_pipeline.predict(input_data)