Update app.py

app.py

@@ -156,29 +156,63 @@ def create_time_series_plot(metric, start_date, end_date):
 
 def predict_crime_level(crime_felony, crime_misd, crime_viol, sr311_total, dob_permits_total):
     """Predicts crime level based on input features."""
+    print(f"DEBUG: predict_crime_level called with inputs: {crime_felony}, {crime_misd}, {crime_viol}, {sr311_total}, {dob_permits_total}")
+    
     if model is None:
-        return "Model not loaded. Please train and provide a model file.", {}
+        # Create a dummy prediction based on simple logic when model is not available
+        total_crime = crime_felony + crime_misd + crime_viol
+        
+        # Simple rule-based classification for demonstration
+        if total_crime <= 20:
+            prediction = "Low"
+            confidence = {"Low": 0.7, "Medium": 0.2, "High": 0.1}
+        elif total_crime <= 50:
+            prediction = "Medium"
+            confidence = {"Low": 0.2, "Medium": 0.6, "High": 0.2}
+        else:
+            prediction = "High" 
+            confidence = {"Low": 0.1, "Medium": 0.3, "High": 0.6}
+        
+        # Factor in 311 requests and permits
+        if sr311_total > 500:
+            # High service requests might indicate more issues
+            if prediction == "Low":
+                prediction = "Medium"
+                confidence = {"Low": 0.4, "Medium": 0.5, "High": 0.1}
+        
+        if dob_permits_total > 25:
+            # High construction activity might indicate development/change
+            confidence["Medium"] = min(0.8, confidence.get("Medium", 0) + 0.2)
+            
+        print(f"DEBUG: Dummy prediction result: {prediction}, confidence: {confidence}")
+        return f"Predicted Crime Level: {prediction} (using fallback model)", confidence
 
-    # Create a DataFrame for the model
-    input_data = pd.DataFrame({
-        'crime_felony': [crime_felony],
-        'crime_misd': [crime_misd],
-        'crime_viol': [crime_viol],
-        'sr311_total': [sr311_total],
-        'dob_permits_total': [dob_permits_total]
-    })
-    
-    # Predict probabilities
-    probabilities = model.predict_proba(input_data)[0]
-    labels = model.classes_
-    
-    # Get the prediction
-    prediction = labels[np.argmax(probabilities)]
-    
-    # Create a confidence dictionary
-    confidence = {label: prob for label, prob in zip(labels, probabilities)}
-    
-    return f"Predicted Crime Level: {prediction}", confidence
+    try:
+        # Create a DataFrame for the model
+        input_data = pd.DataFrame({
+            'crime_felony': [crime_felony],
+            'crime_misd': [crime_misd],
+            'crime_viol': [crime_viol],
+            'sr311_total': [sr311_total],
+            'dob_permits_total': [dob_permits_total]
+        })
+        
+        # Predict probabilities
+        probabilities = model.predict_proba(input_data)[0]
+        labels = model.classes_
+        
+        # Get the prediction
+        prediction = labels[np.argmax(probabilities)]
+        
+        # Create a confidence dictionary
+        confidence = {label: prob for label, prob in zip(labels, probabilities)}
+        
+        print(f"DEBUG: Real model prediction result: {prediction}, confidence: {confidence}")
+        return f"Predicted Crime Level: {prediction}", confidence
+        
+    except Exception as e:
+        print(f"DEBUG: Error in model prediction: {e}")
+        return f"Error in prediction: {str(e)}", {}
 
 def forecast_time_series(geoid):
     """Forecasts crime for a specific GEOID."""
@@ -298,11 +332,11 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
                 gr.Markdown("Adjust the sliders to reflect the current month's data for a census tract.")
                 with gr.Row():
                     with gr.Column():
-                        felony_slider = gr.Slider(0, 100, label="Felony Count", step=1)
-                        misd_slider = gr.Slider(0, 200, label="Misdemeanor Count", step=1)
-                        viol_slider = gr.Slider(0, 200, label="Violation Count", step=1)
-                        sr311_slider = gr.Slider(0, 1000, label="311 Service Requests", step=10)
-                        dob_slider = gr.Slider(0, 50, label="DOB Permits Issued", step=1)
+                        felony_slider = gr.Slider(0, 100, label="Felony Count", step=1, value=5)
+                        misd_slider = gr.Slider(0, 200, label="Misdemeanor Count", step=1, value=15)
+                        viol_slider = gr.Slider(0, 200, label="Violation Count", step=1, value=10)
+                        sr311_slider = gr.Slider(0, 1000, label="311 Service Requests", step=10, value=100)
+                        dob_slider = gr.Slider(0, 50, label="DOB Permits Issued", step=1, value=3)
                         predict_button = gr.Button("Predict")
                     with gr.Column():
                         prediction_output = gr.Label(label="Prediction Result")