# google/differential-privacy

Google's differential privacy libraries.

Repository: https://github.com/google/differential-privacy
Canonical: https://ross.abutalabs.com/products/differential-privacy
Language: Go
License: Apache-2.0
License Family: permissive
Topics: privacy, differential-privacy, anonymization
Last push: 2026-08-24T21:03:12+00:00

## Health v2 (maintenance only)
Score: 78/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 37, longevity 100
- inputs: {"age_days": 2555, "days_push": 9, "days_rel": 208, "gap_med": 274, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3346, forks 430 (observed 2026-08-28T04:07:57.641309+00:00)

## What it is
Google's collection of differential privacy libraries for computing ε- and (ε, δ)-differentially private statistics over datasets. It includes end-to-end frameworks (Privacy on Beam for Go, PipelineDP4j for JVM), low-level noise/aggregation building blocks in C++, Go, and Java, a privacy budget accounting library, a DP auditing tool, and a CLI for differentially private SQL queries.

## Use cases
- compute differentially private statistics over a dataset
- add calibrated noise to aggregations to protect user privacy
- track and account for privacy budget (epsilon) in a data pipeline
- run differentially private SQL queries
- audit whether a system's differential privacy guarantees hold
- build anonymized analytics on Apache Beam or Spark pipelines

## When to choose
- you need production-grade differential privacy in Go, Java/Kotlin/Scala, C++, or Python
- you want an easy end-to-end DP framework on top of Beam or Spark
- you need low-level DP primitives like noise addition and private aggregations
- you must track privacy budget or audit DP guarantees

## When to avoid
- you need simple anonymization or pseudonymization rather than formal DP guarantees
- you need DP support in a language not covered (e.g., Ruby, PHP)
- you rely on the experimental tools (SQL CLI, auditor) needing API stability
- your team has no familiarity with differential privacy concepts and cannot invest in learning them

## Facets
- artifact type: library
- maturity: active
- function: privacy, security, data-science, testing, cli
- domain: privacy, data-science, analytics, security, developer-tools
- platform: go, cpp, jvm, python, cross-platform
- tags: differential-privacy, anonymization, apache-beam, apache-spark, privacy-budget-accounting, sql, statistics

## Member repositories
- google/differential-privacy (main) score 78

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:57.641309+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:41:00.855059+00:00, confidence not recorded.
  - readme: https://github.com/google/differential-privacy (fetched 2026-08-28T04:07:57.641309+00:00, sha 7b6019cccfc1)
- Data as of 2026-08-30T08:39:29.467469+00:00.