Upload 3 files

README.md

@@ -1,12 +1,52 @@
 ---
-title: ApartmentPricePredictor
-emoji: 🏆
-colorFrom: red
-colorTo: yellow
+title: Zürich Apartment Price Predictor
+emoji: 🏠
+colorFrom: blue
+colorTo: red
 sdk: gradio
-sdk_version: 6.9.0
+sdk_version: 6.8.0
 app_file: app.py
 pinned: false
+short_description: Machine Learning Apartment Rent Price Predictor for Zurich
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# Model Iterations Documentation
+## Task: Apartment Price Prediction (Regression)
+
+---
+
+---
+
+## Summary of Iterative Process
+
+| Iteration | Objective | Key Changes | Models Used | CV Mean R² | CV Std Dev | Test MAE (CHF) | Fit Diagnosis |
+|-----------|-----------|-------------|-------------|------------|------------|----------------|---------------|
+| **1** | Build baseline model | - Basic cleaning<br>- 7 numerical features<br>- Train-test split (80/20)<br>- 5-fold CV | Linear Regression<br>Random Forest (n_estimators=100) | 0.429 (LR)<br>0.441 (RF) | 0.073 (LR)<br>0.075 (RF) | 549.90 (LR)<br>553.98 (RF) | ☐ Overfitting ☑ Underfitting ☐ Good Fit |
+| **2** | Improve through feature engineering | - 7 new features created<br>- Feature scaling (Ridge)<br>- Hyperparameter tuning<br>- 5-fold CV | Ridge (alpha=10.0)<br>Random Forest (n_estimators=200, max_depth=15)<br>Gradient Boosting (n_estimators=150, max_depth=5) | 0.461 (Ridge)<br>0.481 (RF)<br>0.521 (GB) | 0.073 (Ridge)<br>0.070 (RF)<br>0.068 (GB) | 541.11 (Ridge)<br>544.92 (RF)<br>537.78 (GB) | ☐ Overfitting ☐ Underfitting ☑ Good Fit |
+
+---
+
+## Notes
+
+**Metric:** MAE (Mean Absolute Error), R² (5-Fold Cross-Validation)
+
+**Created Features:**
+- rooms_per_sqm: Raumdichte (Zimmer pro m²)
+- wealth_index: Wohlstandsindikator (tax_income × emp normalisiert)
+- is_zurich_city: Binär-Feature für Zürich Stadt (PLZ 8000-8099)
+- pop_emp_ratio: Bevölkerungs-Arbeitsplatz-Verhältnis
+- log_area: Log-Transformation der Wohnfläche
+- log_pop: Log-Transformation der Bevölkerung
+- log_tax_income: Log-Transformation des Steuereinkommens
+
+**Final Selected Features:**
+- area (28.7% importance)
+- log_area (25.0% importance)
+- is_zurich_city (13.0% importance)
+- rooms_per_sqm (12.2% importance)
+- rooms, log_pop, pop_dens, pop, pop_emp_ratio, log_tax_income, emp, wealth_index, tax_income, frg_pct
+
+**Final Model:** Gradient Boosting Regressor
+
+**Reason for Selection:**
+Best R² score (0.521) and lowest MAE (537.78 CHF). Consistent performance between CV and test set with lowest standard deviation (±67.62 CHF), indicating stable predictions.

app.py

@@ -0,0 +1,135 @@
+import gradio as gr
+import pandas as pd
+import numpy as np
+import pickle
+import os
+
+# Load model (or create dummy if not exists)
+MODEL_PATH = "models/model.pkl"
+
+if os.path.exists(MODEL_PATH):
+    with open(MODEL_PATH, 'rb') as f:
+        model_package = pickle.load(f)
+    
+    # Check if it's the new model format (dict) or old format (single model)
+    if isinstance(model_package, dict):
+        model = model_package['model']
+        scaler = model_package.get('scaler', None)
+        features = model_package.get('features', None)
+        model_type = model_package.get('model_type', 'unknown')
+        using_dummy = False
+        print(f"✅ Loaded model: {model_type}")
+    else:
+        # Old format - single model
+        model = model_package
+        scaler = None
+        features = None
+        model_type = 'legacy'
+        using_dummy = False
+        print("✅ Loaded legacy model")
+else:
+    # Dummy model for initial deployment
+    model = None
+    scaler = None
+    features = None
+    model_type = None
+    using_dummy = True
+    print("⚠️ No model found - using dummy")
+
+def create_features(rooms, area, postalcode, pop, pop_dens, frg_pct, emp, tax_income):
+    """Create engineered features from input"""
+    data = {
+        'rooms': rooms,
+        'area': area,
+        'pop': pop,
+        'pop_dens': pop_dens,
+        'frg_pct': frg_pct,
+        'emp': emp,
+        'tax_income': tax_income,
+        # Engineered features
+        'rooms_per_sqm': rooms / area,
+        'wealth_index': (tax_income / 100000) * (emp / 100000),
+        'is_zurich_city': 1 if (postalcode >= 8000 and postalcode < 8100) else 0,
+        'pop_emp_ratio': pop / (emp + 1),
+        'log_area': np.log1p(area),
+        'log_pop': np.log1p(pop),
+        'log_tax_income': np.log1p(tax_income)
+    }
+    return pd.DataFrame([data])
+
+def predict_price(rooms, area, postalcode, pop, pop_dens, frg_pct, emp, tax_income):
+    """
+    Predict apartment rental price based on input features
+    """
+    if using_dummy:
+        # Simple dummy calculation for testing
+        base_price = area * 25  # rough estimate: 25 CHF per m²
+        room_factor = rooms * 200
+        location_factor = (tax_income / 1000) * 0.5
+        
+        estimated_price = base_price + room_factor + location_factor
+        return f"💰 Geschätzter Mietpreis: CHF {estimated_price:.2f}/Monat\n\n⚠️ Dummy-Modell - Training folgt!"
+    
+    else:
+        # Create feature dataframe
+        input_data = create_features(rooms, area, postalcode, pop, pop_dens, frg_pct, emp, tax_income)
+        
+        # Select only the features the model was trained on
+        if features is not None:
+            input_data = input_data[features]
+        
+        # Apply scaling if needed (for Ridge)
+        if scaler is not None:
+            input_data_scaled = scaler.transform(input_data)
+            prediction = model.predict(input_data_scaled)[0]
+        else:
+            prediction = model.predict(input_data)[0]
+        
+        return f"💰 Geschätzter Mietpreis: CHF {prediction:.2f}/Monat\n\n🤖 Modell: {model_type.replace('_', ' ').title()}"
+
+# Gradio Interface
+with gr.Blocks(title="Zürich Apartment Price Predictor") as demo:
+    gr.Markdown("# 🏠 Zürich Apartment Rent Predictor")
+    gr.Markdown("Vorhersage von Mietpreisen für Wohnungen im Kanton Zürich")
+    
+    with gr.Row():
+        with gr.Column():
+            rooms = gr.Number(label="Anzahl Zimmer", value=3.5, minimum=1, maximum=10)
+            area = gr.Number(label="Wohnfläche (m²)", value=75, minimum=10, maximum=500)
+            postalcode = gr.Number(label="Postleitzahl", value=8001, minimum=8000, maximum=8999)
+            
+        with gr.Column():
+            pop = gr.Number(label="Bevölkerung (Gemeinde)", value=420217)
+            pop_dens = gr.Number(label="Bevölkerungsdichte", value=4778.99)
+            frg_pct = gr.Number(label="Ausländeranteil (%)", value=32.46)
+            emp = gr.Number(label="Anzahl Arbeitsplätze", value=491193)
+            tax_income = gr.Number(label="Durchschn. Steuereinkommen", value=85446)
+    
+    predict_btn = gr.Button("🔍 Preis berechnen", variant="primary")
+    output = gr.Textbox(label="Ergebnis", lines=3)
+    
+    predict_btn.click(
+        fn=predict_price,
+        inputs=[rooms, area, postalcode, pop, pop_dens, frg_pct, emp, tax_income],
+        outputs=output
+    )
+    
+    gr.Markdown("---")
+    gr.Markdown("### 📊 Beispiel-Werte")
+    gr.Markdown("""
+    - **Zürich City (8001)**: PLZ 8001, Pop: 420217, Pop_dens: 4778.99, Frg_pct: 32.46, Emp: 491193, Tax: 85446
+    - **Winterthur (8400)**: PLZ 8400, Pop: 114220, Pop_dens: 2009.6, Frg_pct: 29.9, Emp: 57583, Tax: 72190
+    - **Rüti ZH (8630)**: PLZ 8630, Pop: 12286, Pop_dens: 1221.3, Frg_pct: 24.8, Emp: 5053, Tax: 66676
+    """)
+    
+    gr.Markdown("---")
+    gr.Markdown("### 🎯 Modell-Performance")
+    gr.Markdown("""
+    - **Model**: Gradient Boosting Regressor
+    - **R² Score**: 0.521
+    - **MAE**: 537.78 CHF
+    - **Features**: 14 (inkl. engineered features)
+    """)
+
+if __name__ == "__main__":
+    demo.launch(theme=gr.themes.Soft())

requirements.txt

@@ -0,0 +1,7 @@
+gradio>=4.0.0
+pandas>=2.0.0
+scikit-learn>=1.3.0
+numpy>=1.24.0
+matplotlib>=3.7.0
+seaborn>=0.12.0
+xgboost>=2.0.0