DCAT-US metadata

DCAT-US (also called data.json) is the US government's profile of DCAT - the schema every federal agency publishes so the central catalog can harvest it. A DCAT-US record describes one dataset and links its file-level distributions.

The cataloged records show the standard from several angles. The Data.gov Telecommunications Keyword Search lists 'DCAT-US metadata' in data_formats and states that 'each dataset's own license field governs reuse', with the licence 'carried in each record's DCAT license field'; its granularity is 'Dataset-level metadata records with linked resource distributions', tagged 'dcat-us' and 'metadata-search'. Another record describes 'DCAT-US JSON metadata' as its short form. The health insurance search returns results 'each with DCAT metadata, download formats and harvest provenance'; the Medicaid portal lists 'DCAT-US data.json' among its formats; and the Cell Tower Datasets Catalog notes that 'Data.gov aggregates agency DCAT-US listings', with licensing set per publisher (Connecticut commercial delivery terms, Austin under city open-data policy).

DCAT-US standardizes the description; it deliberately does not standardize the rights, and that split drives buyer risk.

• The license field is the whole rights story. Because 'each dataset's own license field governs reuse', a harvest that ignores the field is redistributing on assumption - and some records leave it empty.

• Publishers differ inside one catalog. The cataloged cell-tower slice mixes Connecticut public-domain records with City of Austin files under the city's own open-data policy; one catalogue, several regimes.

• data.json is the machine entry point. The Medicaid portal lists 'DCAT-US data.json' as a format, meaning agencies publish the schema directly - a stable target for scheduled harvests.

• Harvest provenance travels with the record. Fields like download formats and harvest provenance let you judge whether a record is fresh and where it came from before you rely on it.

1. Read the DCAT license field first. The cataloged rule is that 'each dataset's own license field governs reuse' and the licence is 'carried in each record's DCAT license field'.
2. Confirm the record exposes distributions, not just a landing page. Granularity is 'Dataset-level metadata records with linked resource distributions'; distribution links are what make automated retrieval possible.
3. Check for data.json availability at the agency. The Medicaid portal lists 'DCAT-US data.json' among its formats.

Access and licence terms that sit alongside DCAT-US:

• api-data-gov-key-rate-limits — The keyed alternative to unauthenticated DCAT-US search, with its own quota
• cell-tower — A query slice showing agency DCAT-US listings with publisher-set licences.
• cursor-paginated-search-api — How the catalog's DCAT-US results are paged for harvesting.
• managed care organization — A managed-health-care use case served by DCAT-described federal records.