sedegah/eta

A comprehensive machine learning system for predicting traffic speeds and calculating estimated time of arrival (ETA) in Accra, Ghana. # Accra Traffic Prediction & ETA Engine A comprehensive machine learning system for predicting traffic speeds and calculating estimated time of arrival (ETA) in Accra, Ghana. The engine uses XGBoost models trained on historical traffic data, weather conditions, and event information to provide accurate travel time predictions. ## Features - **Real-time Traffic Speed Prediction**: ML-powered speed predictions for different road segments - **Multi-Route ETA Calculation**: Calculate and compare ETAs for multiple route options - **Route Optimization**: Find the fastest route based on current conditions - **Traffic Pattern Analysis**: Analyze historical traffic patterns and trends - **Departure Time Optimization**: Determine the best time to travel - **Weather Integration**: Factor in weather conditions (rain, temperature, humidity) - **Event Awareness**: Account for special events that impact traffic - **Prediction Caching**: Improved performance for repeated queries - **Confidence Intervals**: Uncertainty estimates for all predictions ## Project Structure ``` eta/ ├── eta.ipynb # Main Jupyter notebook with the complete implementation ├── README.md # Project documentation (this file) └── data/ ├── traffic_data.csv # Historical traffic speed data ├── weather_data.csv # Weather conditions data └── events_data.csv # Special events data ``` ## Getting Started ### Prerequisites - Python 3.8+ - Jupyter Notebook or JupyterLab ### Installation 1. Clone the repository: ```bash git clone github.com cd eta ``` 2. Install required dependencies: ```bash pip install -r requirements.txt ``` 3. Launch Jupyter Notebook: ```bash jupyter notebook eta.ipynb ``` ## Usage ### Validate Data Before Training Run the built-in validator to confirm the CSV files are complete and internally consistent: ```bash python scripts/validate_data.py ``` This checks schema/order correctness, timestamp format validity, value ranges, kn …