DCAT-AP metadata

DCAT-AP is the EU's profile of the W3C Data Catalogue Vocabulary: a fixed set of properties - title, description, publisher, licence, distribution - used to describe a dataset as a catalogue record. A DCAT-AP record is metadata about a dataset, not the dataset itself.

The cataloged records show the profile operating at scale. data.europa.eu - EU Open Data Portal is tagged 'dcat-ap' and 'metadata aggregation', with granularity described as 'Dataset-level metadata records'; the aluminium search slice reports roughly 550 matching records out of about 1.7 million catalogued datasets. The wholesale-trade record describes the shape precisely: 'One DCAT-AP metadata record per dataset with per-distribution detail'. Licences ride on distributions.

DCAT-AP tells you where to look, not what you will find, and treating the record as the data is the classic aggregator mistake.

• Record count is not data volume. The aluminium query returns about 550 records; each points at a different source with its own granularity, from point registers to network polylines.

• Licences live per distribution. Because the licence URI is attached to each distribution, a dataset can mix open and restricted files; harvesting at dataset level risks redistributing a file whose licence you never read.

• Harvest lag is real. data.europa.eu re-publishes metadata harvested from national portals, so the record can be current while the underlying series is not.

• The profile still pays off. One DCAT-AP parser reads records from 36 countries' catalogues instead of writing a scraper per portal, which is exactly why the aggregation layer exists.

1. Confirm records are DCAT-AP, not bespoke JSON. The cataloged portal records carry tags 'dcat-ap' and formats 'JSON-LD / DCAT-AP'; a source without the profile means per-portal parsing work.
2. Read the licence at distribution level. data.europa.eu declares 'licence per distribution via a URI (e.g. creativecommons.org/licenses/by/4.0)'.
3. Follow the record to the source. Granularity is 'Dataset-level metadata records'; the underlying content 'ranges from point registers to network polylines', so only the source page tells you what you actually get.

Metadata-layer terms that sit next to DCAT-AP:

• commercial delivery terms — One of the open licences declared per distribution on harvested EU portal records.
• commercial delivery terms license — The rights position when a harvested record declares no licence restrictions at all
• mineral commodity summaries — Related term in the Datadory data glossary.
• combined ratio — Related term in the Datadory data glossary.