Datapakken

Datapakken, or data package, is a portable, self-describing bundle for distributing data. It originated in the open data community as part of the Frictionless Data initiative led by the Open Knowledge Foundation, with the aim of making datasets easily shareable and interoperable across different systems. A datapakken is designed so software can understand its contents without additional context.

A datapakken consists of a root descriptor file named datapackage.json and one or more data resources. The descriptor holds metadata such as name, title, description, license, keywords, and sources, along with a resources array. Each resource describes a data file, including name, path, format, and mediatype, and may include a schema that defines the structure of the data (for example table columns and types). Common resource formats include CSV, JSON, and Excel. The schema helps enforce data quality and consistency during ingestion and validation.

Creating a datapakken involves collecting data files, writing the datapackage.json descriptor to describe them, and optionally validating the contents against the declared schema with tools such as GoodTables. The complete package can be compressed into a ZIP file for distribution and then loaded by data portals or applications that support the Data Package standard.

Datapakken enables portability, consistent metadata, and easier data discovery across platforms. It is widely used in open data portals, data catalogs, and data science workflows as a lightweight alternative to larger, more rigid data formats. See also: Data Package specification, Frictionless Data, GoodTables. Example of a minimal datapackage.json: { "name": "example-dataset", "resources": [ { "name": "people", "path": "data/people.csv", "format": "csv", "schema": { "fields": [ { "name": "id", "type": "integer" }, { "name": "name", "type": "string" } ] } } ] }