W3C DCAT (Data Catalog Vocabulary)

DCAT (Data Catalog Vocabulary) is a W3C Recommendation: an RDF vocabulary, published under the namespace http://www.w3.org/ns/dcat#, for describing datasets, dataset series, distributions, and data services within a data catalog so that catalogs published by different organizations can be harvested, aggregated, and searched interoperably. A metadata record qualifies as DCAT (rather than a merely similar, ad hoc catalog schema) when its fields map to DCAT's own RDF classes and properties -- dcat:Catalog, dcat:Dataset, dcat:Distribution, dcat:DataService, dcat:DatasetSeries, and dcat:CatalogRecord, built on top of Dublin Core, FOAF, SKOS, and PROV-O terms rather than a locally invented field set -- and is expressed (or losslessly convertible to) RDF, typically serialized as JSON-LD, Turtle, or RDF/XML.

ByCASRAI Editorial Board

· Last updated 18 Jul 2026

Worked examples

Is an instance

A national open-data portal (e.g. data.europa.eu, built on DCAT-AP) exposes each dataset as a dcat:Dataset with dct:title, dct:description, and dcat:keyword, linked to one or more dcat:Distribution resources -- a CSV download, a JSON API endpoint -- each carrying its own dcat:accessURL and dct:format, all wrapped in a dcat:Catalog resource that a harvester can crawl to aggregate metadata across many national and regional portals into one federated search index.

Is an instance

A U.S. federal agency's data.json file, required under the U.S. Project Open Data policy, is itself a dcat:Catalog instance conforming to DCAT-US (currently v3.0), a U.S.-specific application profile of W3C DCAT 3 published via resources.data.gov; Data.gov harvests these per-agency catalogs to build its own aggregate catalog.

Looks similar, but isn't

Not an instance

A university library's plain HTML dataset listing page, with no RDF serialization, no dcat: namespace terms, and no machine-readable catalog file, may describe datasets in a broadly similar way but is not DCAT -- DCAT requires the description to actually be expressed in (or convertible to) the vocabulary's own RDF classes and properties, not merely to cover similar ground in prose.

Not an instance

A repository's Dublin Core or DataCite metadata record for a single dataset is not itself DCAT, even though DCAT reuses Dublin Core terms (dct:title, dct:description) for individual properties -- DCAT is specifically a catalog-level vocabulary, with dcat:Catalog and dcat:Dataset as first-class RDF classes and relationships between them, not a flat per-item element set.

What DCAT is

DCAT (Data Catalog Vocabulary) is a World Wide Web Consortium (W3C) Recommendation — an RDF vocabulary, in the same technical family as PROV-O and SKOS, purpose-built to let data catalogs published by different organizations describe their holdings in a common, machine-readable structure so the catalogs can be harvested, aggregated, and searched together. Before DCAT, government and institutional data portals each invented their own catalog schema; a harvester wanting to build a cross-portal search index had to write a custom parser per source. DCAT gives every catalog a shared RDF vocabulary — published under the namespace http://www.w3.org/ns/dcat# — so a single generic harvester can consume any DCAT-conformant catalog.

Version history

DCAT (2014) — the original Recommendation, developed by W3C’s Government Linked Data Working Group and published 16 January 2014, aimed primarily at government open-data portals.
DCAT2 (4 February 2020) — a substantial revision by the Dataset Exchange Working Group, adding the dcat:Resource superclass (the common parent of Dataset and the newly added DataService), dcat:DataService for describing APIs, and broader applicability beyond government portals to scientific, statistical, and commercial catalogs.
DCAT3 (22 August 2024) — the current Recommendation, adding dcat:DatasetSeries for grouping related, evolving datasets (e.g. successive releases of a time series), versioning properties, and checksum support for distributions.

All three versions share the same core design and the same namespace; DCAT3 is additive rather than a breaking redesign of DCAT2.

Core classes

dcat:Catalog — a curated collection of metadata about resources (typically datasets and services); a catalog is itself modeled as a dataset, so catalogs can be nested or cross-referenced.
dcat:Resource — the abstract parent class (introduced in DCAT2) shared by Dataset and DataService, carrying common descriptive properties like dct:title, dct:description, dcat:keyword, and dct:publisher.
dcat:Dataset — a collection of data, published or curated by a single agent, and available for access or download in one or more formats.
dcat:Distribution — a specific, accessible representation of a dataset: a downloadable CSV file, a Turtle dump, an API endpoint — each with its own format, access URL, and (as of DCAT3) checksum.
dcat:DataService (DCAT2+) — a site or endpoint providing programmatic access to one or more datasets, such as a REST or SPARQL API.
dcat:DatasetSeries (DCAT3) — a group of datasets sharing a common theme and structure but differing along some dimension, such as a monthly statistical release.
dcat:CatalogRecord — metadata about an entry within a catalog (e.g. when the entry was added or last modified to the catalog itself), distinct from metadata about the resource the entry describes.

DCAT deliberately reuses established vocabularies rather than reinventing them: descriptive properties draw on Dublin Core terms (dct:title, dct:description, dct:license), agents and contacts on FOAF, controlled-vocabulary concepts (themes, categories) on SKOS, and provenance on PROV-O.

Government open-data adoption

DCAT’s first version was written with government open-data portals as the primary use case, and that remains its most visible deployment. The European Data Portal (data.europa.eu) is built on DCAT-AP, an application profile of DCAT maintained through the European Commission’s interoperability programme that constrains and extends the base vocabulary with EU-specific requirements (controlled vocabularies for file types, licences, and spatial/temporal coverage) so that national and regional portals across EU member states can be harvested into one federated catalog. In the United States, the DCAT-US Schema (formerly the Project Open Data Metadata Schema, currently at v3.0 as a U.S. application profile of W3C DCAT3) is the metadata standard U.S. federal agencies use to publish their data.json catalog file under the Project Open Data policy; Data.gov harvests these per-agency catalogs to build its own aggregate federal data catalog. Both profiles follow the same underlying pattern: DCAT provides the shared vocabulary, and a national or regional “application profile” adds the constraints and controlled vocabularies a specific jurisdiction needs on top of it.

References

W3C, “Data Catalog Vocabulary (DCAT) – Version 3,” W3C Recommendation, 22 August 2024 (w3.org/TR/vocab-dcat-3/).
W3C, “Data Catalog Vocabulary (DCAT),” W3C Recommendation, 16 January 2014 (original version, w3.org/TR/vocab-dcat/).
resources.data.gov, “DCAT-US Schema v3.0” and “Data standards: DCAT-US.”

Use in your systems

JATS XML <role> element

xml

<role vocab="credit"
      vocab-identifier="https://casrai.org/dictionary/"
      vocab-term="W3C DCAT (Data Catalog Vocabulary)"
      vocab-term-identifier="https://casrai.org/dictionary/term/dcat-data-catalog-vocabulary" />

Schema.org DefinedTerm (JSON-LD)

json

{
  "@context": "https://schema.org",
  "@type": "DefinedTerm",
  "@id": "https://casrai.org/dictionary/term/dcat-data-catalog-vocabulary",
  "name": "W3C DCAT (Data Catalog Vocabulary)",
  "identifier": "https://casrai.org/dictionary/term/dcat-data-catalog-vocabulary",
  "description": "DCAT (Data Catalog Vocabulary) is a W3C Recommendation: an RDF vocabulary, published under the namespace http://www.w3.org/ns/dcat#, for describing datasets, dataset series, distributions, and data services within a data catalog so that catalogs published by different organizations can be harvested, aggregated, and searched interoperably. A metadata record qualifies as DCAT (rather than a merely similar, ad hoc catalog schema) when its fields map to DCAT's own RDF classes and properties -- dcat:Catalog, dcat:Dataset, dcat:Distribution, dcat:DataService, dcat:DatasetSeries, and dcat:CatalogRecord, built on top of Dublin Core, FOAF, SKOS, and PROV-O terms rather than a locally invented field set -- and is expressed (or losslessly convertible to) RDF, typically serialized as JSON-LD, Turtle, or RDF/XML.",
  "inDefinedTermSet": "https://casrai.org/dictionary/domain/data-infrastructure#set",
  "url": "https://casrai.org/dictionary/term/dcat-data-catalog-vocabulary",
  "sameAs": [],
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "publisher": {
    "@id": "https://casrai.org/#organization"
  },
  "dateModified": "2026-07-18T06:30:48",
  "inLanguage": "en"
}