Welcome to DrivenData's GitHub Page, a home for open source code in support of data science, machine learning, and AI for social good.

DrivenData runs data science competitions and works directly with mission-driven organizations to tackle real-world challenges in areas like health, education, conservation, disaster response, and more. Our open source repositories contain tools we built and maintain as well as competition-winning models and community-driven solutions available for everyone to use, learn from, and contribute to.

DrivenData helps mission-driven organizations harness data to work smarter and deliver greater social impact. We believe in:

• Open collaboration through accessible machine learning, AI, and data science.
• Sharing learning and tools from both our work and our competitions to benefit the global data community.
• Supporting social good, enabling data scientists to solve problems that matter.

Learn more about our work on our website.

We host a variety of open source repositories, including tools for data science workflows, purpose-built packages in specific domains, and winning models and approaches from our competitions.

We open source practical tools we use in our own work to support reproducible, responsible, and maintainable software.

• cookiecutter-data-science: A standardized yet flexible data science project template.
• cloudpathlib: pathlib-style interfaces for cloud storage.
• deon: A CLI tool for adding ethics checklists to data science workflows.
• erdantic: Generate entity relationship diagrams from Python models.

We collaborate with partner organizations to build and deliver open source applications that address domain-specific social impact challenges.

• zamba: A deep learning framework for wildlife camera trap image classification.
• cyfi: A package for detecting harmful algal blooms from satellite imagery.
• scipeds: A "baked data" library for working with higher education data from IPEDS.

We publish winning solutions from past data science competitions under permissive licenses to support learning and reuse. These repositories collect competition submissions spanning topics such as public health, energy forecasting, natural language challenges, and more.

Check out our contribution guidelines in individual repositories for details on how to get involved!

Projects in this organization are released under licenses stated in their individual repositories. Please check each repository for details on licensing and terms of use.

Thank you for exploring DrivenData’s open source! We look forward to what we can build together.