Linked Data is an approach to structuring and sharing data that enables the individual records in a dataset to be uniquely addressable using proven, scalable Web technologies. Linked Data can be used to publish Open Data for public consumption or to share data across departments within an organisation for data integration purposes.
Data integration is supported by different types of linking between datasets, and is underpinned by the core principle that each record has its own unique identifier.
Linked data connects data points together in such a way that it is possible to enter this web of connections from many different starting points. One user may be interested in drilling down by geographic location whereas another wants to group by department; and a third user may be interested in overlaying their own dataset to create a new organising principle for the same data.
Connections between records provides context that can be used to filter, select and refine a subset of records of interest to an end user. The connections between records makes it possible to explore multiple facets of the same entity, and with the global Web of data, those connections don’t need to be limited to the data you hold within your own organisation.
Rather than producing large aggregated reports that are infrequently updated, the Linked Data approach makes it easy to publish collections of smaller connected datasets that are easier to update frequently or even in real-time. The aggregated reports can then be driven by the use of APIs and query to produce timely insights.
The principle API of Linked Data is the same web protocols that are used by web-browsers. This is important because it provides a lowest-common-denominator form of access to your data that relies on nothing more than a web connection. For more advanced use-cases, there is a full query language for Linked Data known as SPARQL.