What is Linked Data?

Linked data turns separate files into a connected web of information. When an organisation publishes its internal data this way, each dataset can be found, browsed and joined to others without someone first building an integration for it.

That matters most in large organisations with thousands of datasets. Instead of a separate ETL pipeline between every pair of systems, each dataset describes itself and links to the others, so a new connection is a matter of adding a link.

Outside the organisation there is a large body of open linked data to draw on: government statistics, scientific databases, geographic data and industry standards. Linking to it adds trusted reference data, such as official place names or standard codes, without having to copy and maintain it.

Linked data is built on three web standards:

URIs. Every entity, whether a person, product, concept or dataset, gets a globally unique identifier that looks like a web address.

HTTP. Those URIs can be looked up. Following one returns structured data about the entity, usually in RDF.

RDF. The data comes back as subject-predicate-object triples, and includes links to other URIs, so you can keep following them from one dataset to the next.

Berners-Lee also proposed a five-star scheme for open data. One star means the data is on the web in any format under an open licence; five stars means it uses URIs and RDF and links to other people’s data. The aim is to get as far up the scale as is practical.

BBC. The BBC built its programme and music sites on linked data, giving every programme, artist and topic a stable URI with machine-readable data behind it, which let different BBC services link to the same things consistently.

Government open data. The EU’s data.europa.eu describes the datasets it catalogues with DCAT, an RDF vocabulary, and offers a SPARQL endpoint, so researchers can search and combine data from different countries.

Libraries and museums. The Library of Congress publishes its subject headings and name authorities as linked data at id.loc.gov, and Europeana does the same for millions of items from European collections, so scholars can trace connections between institutions.