HaajraMumtaz/LahoreTrafficPredictor

---

Essential metadata for Hugging Face

license: apache-2.0 tags: - tabular-regression - traffic-prediction - catboost - lahore datasets: - HaajraMumtaz/LahoreTrafficData metrics: - mae - rmse

Lahore Route Predictor 🚦

This model predicts the congestion_multiplier for various routes in Lahore, Pakistan. It uses a CatBoost regressor trained on historical traffic patterns to estimate how much slower a trip will be compared to free-flow traffic given coordinates, type of day and time of day etc

How to Use

To load and use this model, you need the catboost library installed.

Loading the Model

from catboost import CatBoostRegressor

model = CatBoostRegressor()
model.load_model("model.cbm") # Ensure this file is in your folder

Making a Prediction

The model expects features in the following order: origin_zone, dest_zone, road_type, speed_limit_kmh, num_lanes, is_one_way, distance_km, road_curvature, has_signal, is_construction, weather_condition, day_of_week, time_slot, is_weekend, is_holiday, day_type.

Training & Evaluation

• Algorithm: CatBoost Regressor
• Data Source: https://huggingface.co/datasets/HaajraMumtaz/LahoreTrafficData
• Key Features: Road type, construction status, and weather (especially Smog/Rain).