Taxi Demand Predictor

Overview

The Taxi Demand Predictor is a machine learning application designed to predict taxi demand in a given area using historical data. This project leverages various data science and machine learning techniques to provide accurate demand forecasts, which can be beneficial for taxi companies and ride-sharing services.

Features

• Data Ingestion: Fetches and processes historical taxi demand data.
• Machine Learning Models: Utilizes models such as LightGBM and XGBoost for demand prediction.
• Interactive Visualization: Provides a user-friendly interface using Streamlit for visualizing predictions and trends.
• Geospatial Analysis: Integrates geospatial data to enhance prediction accuracy based on location.
• Feature Store Integration: Uses Hopsworks for managing and serving features for model training and inference.

Technologies Used

• Python: The primary programming language for data processing and model development.
• Streamlit: A framework for building interactive web applications.
• Pandas: For data manipulation and analysis.
• Scikit-learn: For implementing machine learning algorithms.
• LightGBM: For gradient boosting framework that uses tree-based learning algorithms.
• XGBoost: An optimized distributed gradient boosting library.
• Geopandas: For geospatial data processing.
• Hopsworks: For managing feature stores and serving features to models.

Installation

To set up the project locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/yourusername/taxi_demand_predictor.git
   cd taxi_demand_predictor

2. Install Poetry if you haven't already:

   curl -sSL https://install.python-poetry.org | python3 -

3. Install the project dependencies:

   poetry install

4. Create a .env file in the project root and add your HOPSWORKS_API_KEY:

   HOPSWORKS_API_KEY="your_api_key_here"
   

Usage

To run the application, use the following command:

poetry run streamlit run src/frontend.py

Open your web browser and navigate to http://localhost:8501 to access the application.

Contributing

Contributions are welcome! If you have suggestions for improvements or new features, please open an issue or submit a pull request.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Acknowledgments

• Hopsworks for providing the feature store.
• Streamlit for enabling easy web app development.
• The open-source community for their contributions and support.