List of implementations of differentially private analyses


Since the advent of differential privacy, a number of systems supporting differentially private data analyses have been implemented and deployed. This article tracks real-world deployments, production software packages, and research prototypes.

Real-world deployments

NameOrganizationYear IntroducedNotesStill in use?
OnTheMap: Interactive tool for exploration of US income and commute patterns.US Census Bureau2008First deployment of differential privacyYes
RAPPOR in Chrome Browser to collect security metricsGoogle2014First widespread use of local differential privacyNo
Emoji analytics; analytics. Improve: QuickType, emoji; Spotlight deep link suggestions; Lookup Hints in Notes. Emoji suggestions, health type usage estimates, Safari energy drain statistics, Autoplay intent detection Apple2017Yes
Application telemetryMicrosoft2017Application usage statistics Microsoft Windows.yes
Flex: A SQL-based system developed for internal Uber analyticsUber2017Unknown
2020 CensusUS Census Bureau2018Yes
Audience Engagement APILinkedIn2020Yes
Labor Market InsightsLinkedIn2020Yes
COVID-19 Community Mobility ReportsGoogle2020Unknown
Advertiser QueriesLinkedIn2020-
U.S. Broadband Coverage Data SetMicrosoft2021Unknown
IRS and Dept. of Education2021Unknown
Ohm ConnectRecurve2021
Israeli Ministry of Health2024Yes

Production software packages

These software packages purport to be usable in production systems. They are split in two categories: those focused on answering statistical queries with differential privacy, and those focused on training machine learning models with differential privacy.

Statistical analyses

NameDeveloperYear IntroducedNotesStill maintained?
Google's differential privacy librariesGoogle2019Building block libraries in Go, C++, and Java; end-to-end framework in Go,.Yes
OpenDPHarvard, Microsoft2020Core library in Rust, SDK in Python with an SQL interface.Yes
Tumult AnalyticsTumult Labs2022Python library, running on Apache Spark.Yes
PipelineDPGoogle, OpenMined2022Python library, running on Apache Spark, Apache Beam, or locally.Yes
PSI : A Private data Sharing InterfaceHarvard University Privacy Tools Project.2016No
TopDown AlgorithmUnited States Census Bureau2020Production code used in the 2020 US Census.No

Machine learning

NameDeveloperYear IntroducedNotesStill maintained?
DiffprivlibIBM2019Python library.Yes
TensorFlow PrivacyGoogle2019Differentially private training in TensorFlow.Yes
OpacusMeta2020Differentially private training in PyTorch.Yes