Linked Data

From Data.gov.uk

Linked data is a term originally coined by Sir Tim Berners-Lee to describe how to publish data on the web by working with the RESTful architecture of the web.

There are four rules to linked data:

1. Use URIs as names for things means that as well as having URIs for documents and dumps of data, we also use URIs for real-world things such as individual schools, roads and local authorities and abstract things such as the concept of a car or a sixth-form college.
2. Use HTTP URIs so that people can look up those names means that the URIs that we use can be put into a browser, accessed through AJAX or downloaded using curl. The information is on the web: you don't need to go to a portal that understands a specific naming scheme to find it, just access the URI.
3. When someone looks up a URI, provide useful information, using the standards (RDF, SPARQL) means that when you access the URI, you should get some machine-readable information back. This rule specifically mentions the standards RDF, which is a way of modelling data, and SPARQL, which is a way of querying data.
4. Include links to other URIs so that they can discover more things means that the data you get about a thing should include links to other related things, creating a web of data.

For people using government data, linked data provides transactional access to the data that you need. Instead of downloading an entire dataset, which you then have to search to get hold of information and regularly keep up to date, you can look up information about the particular item that you're interested in, when you're interested in it. What's more, because the data is linked, it is easy to locate extra information when you need it, rather than having to hunt for yet more datasets that hold the related information.

For example, we have linked data for schools, based on information from EduBase. Within this set of linked data, each school is named using an HTTP URI. For example:

http://education.data.gov.uk/id/school/103335

is the URI that names Nonsuch Primary School in Birmingham. If you click on that link in a browser then you will be taken to an HTML page that describes the school. The HTML page isn't particularly pretty at the moment -- it's just a default view of the data that we have about the school -- but the real value is the machine-readable information.

You can see the machine-readable information by clicking on the links at the bottom of the page to:

http://education.data.gov.uk/doc/school/103335.rdf (RDF/XML)
http://education.data.gov.uk/doc/school/103335.ttl (Turtle)
http://education.data.gov.uk/doc/school/103335.json (RDF/JSON)

These three pages are all representations of the same RDF graph using different syntax.

The information that's available about the school includes links to other, related things, such as the Local learning and skills council that the school falls under (Birmingham and Solihull -- http://statistics.data.gov.uk/id/llsc/WM410), the ward that it's in (Bartley Green -- http://statistics.data.gov.uk/id/local-authority-ward/00CNGU) and the fact that it's a community school (http://education.data.gov.uk/def/school/TypeOfEstablishment_Community_School). These pages themselves contain further information, published in the same way.

Further Resources

• Linked Data Resources
• W3C page on Linked Data
• public-lod@w3.org mailing list