Data Contracts
summary
Claims that data contracts act like an API for datasets, specified in YAML and now standardized as the Open Data Contract Standard, and that they are executable: the Data Contract CLI generates SQL checks from contract fields and runs them against BigQuery/Snowflake/etc to enforce schema, semantics and quality in CI pipelines. Demonstrations include an Excel→YAML workflow, a marketplace (Entropy Data) that applies contracts to access with an 80% consent example, and an MCP proxy that builds SQL from metadata and blocks disallowed uses.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 3:15What is a data contract
- 7:56Data-first vs contract-first
- 10:06Data ownership and requirements
- 13:12Schema and semantics
- 16:24Data quality and limits
- 18:40Service levels and guarantees
- 21:05Tools for contracts
- 23:15Not legal contracts
- 24:45Improve communication and trust
- 26:52Testing with CLI and automation
- 27:55Metadata for marketplaces
- 29:44AI and LLM integration
- 35:52Data governance and compliance
- 37:38Summary
- 38:42Q&A
- 39:23Brownfield adoption
- 40:05Keeping contracts updated
- 40:27Strictness and testing
- 40:48Ownership for external providers
- 41:51Real-time teams
- 42:57Overhead and necessity
- 43:59Ownership across teams
- 45:03Comparison to Pact
- 45:25Business vs technical ownership
- 46:06Changes in ownership
- 46:27Data contracts with manager
related talks
32mData Contracts: Consensus as Code
39mTaming the Chaos: Open Standards for Data Contracts and Products Are the Back...
44mBring data teams together with data contracts
19mUnlock API and Data – The Power of Data Model and Governance
39mData Mesh in the wild
42mDriving Data Democratization
51mWhat is a Data mesh?
35mApplying Data Quality Monitoring