datapakke

Datapakke is a lightweight, portable packaging format for sharing datasets. It originated with the Data Package specification developed by the Open Knowledge Foundation and is used in open data communities to make data self-describing and easy to reuse. A datapakke is typically a directory that contains a metadata file called datapackage.json and one or more data files, such as CSV or JSON.

The datapackage.json provides metadata about the dataset, including its title, description, publisher, license, and version, as well as a list of resources. Each resource describes a data file with properties such as name, path or URL, format, and a schema. The schema defines the fields (column names), data types, formats, and any constraints. Optional fields may include keywords, provenance information, and links to related resources.

Benefits of datapakke include being self-describing and machine-readable, which simplifies ingestion by data portals and data pipelines. They are portable across platforms and programming environments and support validation and versioning as data evolves.

Datapakke is widely used in open data practices and can serve as a starting point for publishing datasets. It is typically accessed as a directory or a compressed archive and can be served over HTTP or hosted in a data catalog, allowing consumers to fetch metadata and data files together.