sdv-dev/SDV

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

3572 stars · 424 forks observed · 2026-10-12

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

SDV (Synthetic Data Vault) is a Python library for generating synthetic tabular data using machine learning models ranging from GaussianCopula to CTGAN. It supports single-table, multi-table relational, and sequential data, with evaluation, visualization, anonymization, and constraint features.

Use cases

• generate synthetic tabular data for testing
• create fake relational multi-table datasets
• anonymize sensitive data while preserving patterns
• generate synthetic time series data
• evaluate quality of synthetic datasets
• train a generative model on my database schema

When to choose

• you need realistic synthetic versions of single, multi-table, or sequential tabular data
• you want on-premises data generation with standard CPUs
• you need built-in quality evaluation, anonymization, and business-rule constraints

When to avoid

• you need synthetic images, text, or audio rather than tabular data
• you need to scale to very large, complex interconnected table sets beyond community edition capabilities
• you require a non-Python environment

Facets

library · maturity stable

data-generation machine-learning deep-learning data-science privacy machine-learning data-science privacy developer-tools python cross-platform synthetic-data tabular-data gan ctgan gaussian-copula anonymization time-series multi-table generative-ai

Member repositories

For agents

markdown · JSON · MCP: product_card(name="sdv-dev/SDV")