Software & Datasets

Selected open-source libraries and datasets associated with my work on privacy-preserving and responsible machine learning.

Libraries & Tools

LDP-Toolbox

Python package and interactive dashboard for analyzing, comparing, and visualizing Local Differential Privacy protocols and their utility, privacy, and attackability trade-offs.

Multi-Freq-LDPy

Python library for one-time, multidimensional, longitudinal, and multidimensional-longitudinal frequency estimation under Local Differential Privacy.

BenchmarkDPFair

Reproducible Python framework for benchmarking fairness-aware learning mechanisms on differentially private synthetic tabular data.

Datasets

MS-FIMU

Open longitudinal synthetic mobility dataset generated from an anonymized call-detail-record setting for privacy and mobility research.

Multivariate-Mobility-Paris

Anonymized and aggregated multivariate human-mobility time series from Paris, suitable for forecasting and privacy-preserving learning experiments.