search

opendatadiscovery / awesome-data-catalogs

📙 Awesome Data Catalogs and Observability Platforms.
MIT License
635 stars 50 forks source link

Updates regarding DataHub #8

Closed sspaeti closed 1 year ago

sspaeti commented 1 year ago

_I got some updates on LinkedIn by Shirshanka Das_: Some facts below for some of the dimensions marked as X above:

  1. Specification-based: DataHub's metadata model is declarative and specified via the open-source Pegasus language (with inter-op with JsonSchema and Avro)
  2. ML 1st Citizen: DataHub supports Features, Models, Notebooks as first-class entities in its metadata model. You will find out of the box integration with feature stores like Feast and machine learning infrastructure like AWS SageMaker.
  3. Data Quality: DataHub has native integration with Great Expectations, and dbt tests for surfacing data quality assertions and results of tests.