OpenDP

A community-driven library for building privacy-preserving statistical analysis tools using differential privacy.

OpenDP is a collaborative project between Harvard University and Microsoft that provides a robust framework for statistical disclosure limitation. The library (written in Rust with Python bindings) enables researchers to compute accurate aggregates like means, variances, and histograms while maintaining rigorous mathematical privacy guarantees. By implementing vetted algorithms for sensitivity calibration and noise injection, OpenDP allows organizations to share sensitive datasets without risking individual re-identification.

https://opendp.org