Today, we’re happy to announce that OpenDP is taking over the maintenance and further development of Tumult Labs’ open-source software! Tumult’s libraries are based on a similar programming framework as OpenDP, and set a high bar for correctness and robustness, so joining efforts is a natural fit. A few former employees of Tumult Labs have joined the OpenDP community and shipped the first release of Tumult Analytics from its new home.
Tumult Analytics is a Python library for differential privacy, focused on scalability, usability, and robustness. It was used by organizations such as the U.S. Census Bureau, the Wikimedia Foundation, and the Internal Revenue Service to publish insights from sensitive data with differential privacy.
You can check out the source code of Tumult Analytics and Tumult Core on OpenDP’s GitHub organization, where we set up automated linting, testing, and a release process; their documentation is now also hosted on OpenDP infrastructure. While we still have more work planned to improve the maintainability of the two libraries, we are now set up to accept contributions and add new features.
Our new release – Tumult Analytics 0.20.2 – does not include any new functionality, but comes with revamped documentation, and confirms that our new release process works.
As always, you can join the OpenDP Slack workspace and post in #lib-support if you have any questions, are looking for help, want to share ideas, or help contribute! You can also subscribe to the OpenDP mailing-list for ongoing updates.
If you’re interested in contributing to Tumult Analytics, we’d love to onboard you and suggest starter projects you could work on! Check out our contribution guidelines, and join the OpenDP Slack to introduce yourself and let us know what you’re interested in.