Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+congestion_multiplier_v1.cbm filter=lfs diff=lfs merge=lfs -text
+training_diagnostics.png filter=lfs diff=lfs merge=lfs -text

congestion_multiplier_v1.cbm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:477a52f47bab33557f78432575b90cf1d5a606084484a34543569156d848f297
+size 3552656

train_traffic_mult.py

@@ -0,0 +1,159 @@
+import os
+import numpy as np
+import pandas as pd
+from catboost import CatBoostRegressor, Pool
+from sklearn.model_selection import train_test_split, KFold
+from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
+import matplotlib.pyplot as plt
+
+# ─── 1. LOAD ─────────────────────────────────────────────────────────────────
+df = pd.read_csv('data/raw/lahore_traffic_dataset.csv')
+print(f"Loaded: {df.shape[0]} rows × {df.shape[1]} cols")
+print(df.dtypes)
+print(df.head(3).to_string())
+
+# ─── 2. VALIDATE ─────────────────────────────────────────────────────────────
+assert df['congestion_multiplier'].between(0.55, 3.5).all(), "Target out of range!"
+assert (df['origin_zone'] != df['dest_zone']).all(), "Self-loops in data!"
+assert df.isnull().sum().sum() == 0, "Null values found!"
+print(f"\n✅ Validation passed")
+print(f"Target range: {df['congestion_multiplier'].min():.3f} – {df['congestion_multiplier'].max():.3f}")
+print(f"Target mean:  {df['congestion_multiplier'].mean():.3f}")
+print(f"Target std:   {df['congestion_multiplier'].std():.3f}")
+
+# ─── 3. FEATURES ─────────────────────────────────────────────────────────────
+TARGET = 'congestion_multiplier'
+
+CAT_FEATURES = [
+    'origin_zone', 'dest_zone', 'road_type',
+    'is_one_way', 'has_signal', 'is_construction',
+    'weather_condition', 'day_of_week', 'time_slot',
+    'is_weekend', 'is_holiday', 'day_type'
+]
+
+NUM_FEATURES = [
+    'speed_limit_kmh', 'num_lanes', 'distance_km', 'road_curvature'
+]
+
+# Cast categoricals to str so CatBoost handles them correctly
+for col in CAT_FEATURES:
+    df[col] = df[col].astype(str)
+
+X = df[CAT_FEATURES + NUM_FEATURES]
+y = df[TARGET]
+
+print(f"\nFeatures: {list(X.columns)}")
+print(f"Cat: {CAT_FEATURES}")
+print(f"Num: {NUM_FEATURES}")
+
+# ─── 4. SPLIT ────────────────────────────────────────────────────────────────
+X_train, X_test, y_train, y_test = train_test_split(
+    X, y, test_size=0.15, random_state=42
+)
+X_train, X_val, y_train, y_val = train_test_split(
+    X_train, y_train, test_size=0.15, random_state=42
+)
+
+print(f"\nTrain: {len(X_train)} | Val: {len(X_val)} | Test: {len(X_test)}")
+
+train_pool = Pool(X_train, y_train, cat_features=CAT_FEATURES)
+val_pool   = Pool(X_val,   y_val,   cat_features=CAT_FEATURES)
+test_pool  = Pool(X_test,  y_test,  cat_features=CAT_FEATURES)
+
+# ─── 5. TRAIN ────────────────────────────────────────────────────────────────
+model = CatBoostRegressor(
+    iterations=2000,
+    learning_rate=0.04,
+    depth=8,
+    l2_leaf_reg=3,
+    min_data_in_leaf=10,
+    loss_function='RMSE',
+    eval_metric='RMSE',
+    random_seed=42,
+    verbose=200,
+)
+
+model.fit(
+    train_pool,
+    eval_set=val_pool,
+    early_stopping_rounds=100,
+)
+
+# ─── 6. EVALUATE ─────────────────────────────────────────────────────────────
+def evaluate(pool, y_true, split_name):
+    preds = model.predict(pool)
+    mae   = mean_absolute_error(y_true, preds)
+    rmse  = np.sqrt(mean_squared_error(y_true, preds))
+    r2    = r2_score(y_true, preds)
+    mape  = np.mean(np.abs((y_true - preds) / y_true)) * 100
+    print(f"\n── {split_name} ──")
+    print(f"  MAE  : {mae:.4f}")
+    print(f"  RMSE : {rmse:.4f}")
+    print(f"  R²   : {r2:.4f}")
+    print(f"  MAPE : {mape:.2f}%")
+    return preds
+
+print("\n=== EVALUATION ===")
+val_preds  = evaluate(val_pool,  y_val,  "Validation")
+test_preds = evaluate(test_pool, y_test, "Test")
+
+# ─── 7. SANITY CHECKS ────────────────────────────────────────────────────────
+print("\n=== SANITY CHECKS ===")
+
+# Check model learned zone personalities
+test_df = X_test.copy()
+test_df['actual'] = y_test.values
+test_df['pred']   = test_preds
+
+print("\nMean predicted multiplier by time_slot:")
+print(test_df.groupby('time_slot')['pred'].mean().sort_values(ascending=False).to_string())
+
+print("\nMean predicted multiplier by road_type:")
+print(test_df.groupby('road_type')['pred'].mean().sort_values(ascending=False).to_string())
+
+print("\nMean predicted multiplier by day_type:")
+print(test_df.groupby('day_type')['pred'].mean().sort_values(ascending=False).to_string())
+
+print("\nMean predicted multiplier by weather_condition:")
+print(test_df.groupby('weather_condition')['pred'].mean().sort_values(ascending=False).to_string())
+
+# ─── 8. FEATURE IMPORTANCE ───────────────────────────────────────────────────
+fi = pd.Series(
+    model.get_feature_importance(),
+    index=CAT_FEATURES + NUM_FEATURES
+).sort_values(ascending=False)
+
+print("\n=== FEATURE IMPORTANCE ===")
+print(fi.to_string())
+
+# ─── 9. PLOTS ────────────────────────────────────────────────────────────────
+fig, axes = plt.subplots(1, 3, figsize=(18, 5))
+
+# Actual vs predicted
+axes[0].scatter(y_test, test_preds, alpha=0.3, s=10)
+axes[0].plot([0.55, 3.5], [0.55, 3.5], 'r--')
+axes[0].set_xlabel('Actual')
+axes[0].set_ylabel('Predicted')
+axes[0].set_title('Actual vs Predicted')
+
+# Residuals
+residuals = y_test.values - test_preds
+axes[1].hist(residuals, bins=50, edgecolor='black')
+axes[1].axvline(0, color='red', linestyle='--')
+axes[1].set_xlabel('Residual')
+axes[1].set_title('Residual Distribution')
+
+# Feature importance
+fi.head(12).plot(kind='barh', ax=axes[2])
+axes[2].invert_yaxis()
+axes[2].set_title('Feature Importance')
+
+plt.tight_layout()
+plt.savefig('models/training_diagnostics.png', dpi=150)
+plt.show()
+
+# ─── 10. SAVE ────────────────────────────────────────────────────────────────
+os.makedirs('models', exist_ok=True)
+model.save_model('models/congestion_multiplier_v1.cbm')
+print("\n✅ Model saved → models/congestion_multiplier_v1.cbm")
+