Update app.py

app.py

@@ -7,82 +7,86 @@ import matplotlib.pyplot as plt
 
 # Load model
 with open("salar_xgb_team.pkl", "rb") as f:
- model = pickle.load(f)
+    model = pickle.load(f)
 
 # Set up SHAP
 explainer = shap.Explainer(model)
 
 # Define prediction function
 def predict_salary(age, education_num, sex, capital_gain, capital_loss, hours_per_week):
- sex_num = 0 if sex == "Male" else 1
- input_data = pd.DataFrame([[age, int(education_num), sex_num, capital_gain, capital_loss, hours_per_week]],
- columns=['age', 'education-num', 'sex', 'capital-gain', 'capital-loss', 'hours-per-week'])
+    sex_num = 0 if sex == "Male" else 1
+    input_data = pd.DataFrame([[age, int(education_num), sex_num, capital_gain, capital_loss, hours_per_week]],
+                              columns=['age', 'education-num', 'sex', 'capital-gain', 'capital-loss', 'hours-per-week'])
 
- pred = model.predict(input_data)[0]
- prob = model.predict_proba(input_data)[0][1]
- label = ">50K" if pred == 1 else "<=50K"
- confidence = f"{prob * 100:.2f}%" if pred == 1 else f"{(1 - prob) * 100:.2f}%"
+    pred = model.predict(input_data)[0]
+    prob = model.predict_proba(input_data)[0][1]
+    label = ">50K" if pred == 1 else "<=50K"
+    confidence = f"{prob * 100:.2f}%" if pred == 1 else f"{(1 - prob) * 100:.2f}%"
 
- # SHAP plot with dark background
- shap_values = explainer(input_data)
- plt.style.use('dark_background')
- fig, ax = plt.subplots(figsize=(6, 3))
- shap.plots.bar(shap_values[0], max_display=6, show=False)
- plt.tight_layout()
+    # SHAP plot with dark background
+    shap_values = explainer(input_data)
+    plt.style.use('dark_background')
+    fig, ax = plt.subplots(figsize=(6, 3))
+    shap.plots.bar(shap_values[0], max_display=6, show=False)
+    plt.tight_layout()
 
- return label, confidence, fig
+    # Conditional message
+    message = "You're halfway to six figures!" if pred == 1 else "Keep working hard!"
+
+    return label, confidence, fig, message
 
 # Custom CSS for styling
 dark_css = """
- body { background-color: #0a0a23; color: white; }
- .gr-box { background-color: #1a1a3d !important; border-radius: 12px; padding: 20px; }
- h1, h3, p { color: white; text-align: center; font-family: 'Segoe UI', sans-serif; }
- .gr-button { background-color: #1db954 !important; color: white !important; border-radius: 10px; font-size: 1.1em; }
- .gr-radio .gr-form { display: flex; justify-content: center !important; }
+    body { background-color: #0a0a23; color: white; }
+    .gr-box { background-color: #1a1a3d !important; border-radius: 12px; padding: 20px; }
+    h1, h3, p { color: white; text-align: center; font-family: 'Segoe UI', sans-serif; }
+    .gr-button { background-color: #1db954 !important; color: white !important; border-radius: 10px; font-size: 1.1em; }
+    .gr-radio .gr-form { display: flex; justify-content: center !important; }
 """
 
 with gr.Blocks(css=dark_css) as demo:
 
- gr.Markdown("""
- 
- 💼 Income Prediction App
- Predict whether someone earns more than $50K/year using financial and demographic data, with AI explainability via SHAP.
- 
- """)
+    gr.Markdown("""
+        <div style='max-width: 700px; margin: 0 auto;'>
+            <h1 style='font-size: 2.5em;'>💼 Income Prediction App</h1>
+            <p style='font-size: 1.2em;'>Predict whether someone earns more than $50K/year using financial and demographic data, with AI explainability via SHAP.</p>
+        </div>
+    """)
 
- with gr.Row():
- with gr.Column():
- gr.Markdown("🎂 Age")
- age = gr.Slider(minimum=0, maximum=100, step=1, value=35, label="", interactive=True)
+    with gr.Row():
+        with gr.Column():
+            gr.Markdown("<h3>🎂 Age</h3>")
+            age = gr.Slider(minimum=0, maximum=100, step=1, value=35, label="", interactive=True)
 
- gr.Markdown("🎓 Education Level")
- education = gr.Dropdown(choices=[str(i) for i in range(1, 17)], value="10", label="", interactive=True)
+            gr.Markdown("<h3>🎓 Education Level</h3>")
+            education = gr.Dropdown(choices=[str(i) for i in range(1, 17)], value="10", label="", interactive=True)
 
- gr.Markdown("🚻 Sex")
- sex = gr.Radio(choices=["Male", "Female"], value="Male", label="", interactive=True)
+            gr.Markdown("<h3>🚻 Sex</h3>")
+            sex = gr.Radio(choices=["Male", "Female"], value="Male", label="", interactive=True)
 
- gr.Markdown("📈 Capital Gain")
- capital_gain = gr.Number(value=0, label="", interactive=True)
+            gr.Markdown("<h3>📈 Capital Gain</h3>")
+            capital_gain = gr.Number(value=0, label="", interactive=True)
 
- gr.Markdown("📉 Capital Loss")
- capital_loss = gr.Number(value=0, label="", interactive=True)
+            gr.Markdown("<h3>📉 Capital Loss</h3>")
+            capital_loss = gr.Number(value=0, label="", interactive=True)
 
- gr.Markdown("🕒 Hours per Week")
- hours_per_week = gr.Number(value=40, label="", interactive=True)
+            gr.Markdown("<h3>🕒 Hours per Week</h3>")
+            hours_per_week = gr.Number(value=40, label="", interactive=True)
 
- predict_btn = gr.Button("🔮 Predict", elem_id="predict-button")
+            predict_btn = gr.Button("🔮 Predict", elem_id="predict-button")
 
- with gr.Row():
- with gr.Column():
- gr.Markdown("📊 Prediction Result")
- result = gr.Textbox(label="", interactive=False)
- confidence = gr.Textbox(label="Confidence", interactive=False)
- shap_plot = gr.Plot(label="SHAP Feature Importance")
+    with gr.Row():
+        with gr.Column():
+            gr.Markdown("<h3>📊 Prediction Result</h3>")
+            result = gr.Textbox(label="", interactive=False)
+            confidence = gr.Textbox(label="Confidence", interactive=False)
+            shap_plot = gr.Plot(label="SHAP Feature Importance")
+            message = gr.Textbox(label="Message", interactive=False)
 
- predict_btn.click(
- fn=predict_salary,
- inputs=[age, education, sex, capital_gain, capital_loss, hours_per_week],
- outputs=[result, confidence, shap_plot]
- )
+    predict_btn.click(
+        fn=predict_salary,
+        inputs=[age, education, sex, capital_gain, capital_loss, hours_per_week],
+        outputs=[result, confidence, shap_plot, message]
+    )
 
 demo.launch()