Skip to content

Linked Data

bandholtz edited this page · 4 revisions

Linked Data

In recent years, Linked Data has evolved into a growing global network of interlinked information which is being maintained by a variety of diverse organizations. Thus the Semantic Web, which was previously considered a largely academic concern, has become a reality. Among the Linked Data participants are the BBC, The New York Times, Wikipedia, MusicBrainz, the German National Library, as well as numerous scientific institutions and governmental bodies. Have times changed? We believe so. It's time for you to join this future - supported by our extensive knowledge and experience regarding semantics and information processing, combined with our REST and Web expertise.

Basic Principles

According to Tim Berners-Lee, inventor of the Web and director of the World Wide Web Consortium (W3C), the Linked Data concept is based on four simple principles:

  • Use URIs as names for things.
  • Use HTTP URIs so that people can look up those names.
  • When someone looks up a URI, provide useful information, using the standards (RDF, SPARQL).
  • Include links to other URIs so that they can discover more things.

This allows both humans and machines to browse and sytematically process interlinked information. Well-designed web applications already provide this opportunity for humans. The RDF and SPARQL standards enable machine-readability, schema-based structure and attribute-specific search.

From the Database Into the Linked Data Network

A simple path into the network of Linked Data is based on your existing (relational) database. The RDB-to-RDF Mapping Language R2RML is used to describe how existing tables, rows and columns are to be represented in RDF. Open source tools like D2R Server take care of the rest; your data becomes available in RDF format and searchable through SPARQL.

The Integrated Web Application

A more elegant way is a web application which combines a human-readable HTML representation of "things" with Linked Data. This requires that the web application is structured like a semantic network. Then all that is required is support for representations matching the capabilities of the client (content negotiation). innoQ's Umweltprobenbank, developed for the German Federal Environment Agency, is one example of such an application. Of course we are happy to develop a similar integrated web application for your needs as well.

RDF Vocabularies

Interoperability of data in the global Linked Data network is greatly simplified if common vocabularies are used for the schema. These include SKOS, Dublin Core, Data Cubes (previously SCOVO), Geonames or Darwin Core. While these do not always define everything just like one might like, RDF's extensibility via sub-classes and sub-properties provides the option to add domain-specific extensions to existing vocabularies. We at innoQ know the standard vocabularies suitable to your needs and formulate extensions as necessary.

Your Corporate Vocabulary

Each enterprise has its own language, which represents the respective subject matter and constitutes an expression of the corporate culture. We can advise you on the formalization of your corporate vocabulary, based on existing glossaries and by analyzing key publications and interviews. This provides you with an RDF schema and a corporate thesaurus which you can maintain and make available on the web, e.g. with our open source solution iQvoc. Examples of this include the German Federal Environment Agency's Semantic Network Service (SNS), which was co-developed by innoQ and is currently being migrated to use iQvoc, as well as Deutschen Bank's SURF project.

Interlinking

With this approach your existing data should already be interconnected internally. What is missing is the link to external data sets like Wikipedia, Geonames or domain-specific databases. Potential partners for interlinking data can be found in the Linked Data catalog.

Here, too, open source is a great asset: We use Silk - "A Link Discovery Framework for the Web of Data" - to semi-automatically create these data links. One example of this is the current Linked Environment Data project.

Become a part of the global Linked Data network and benefit from the improved visibility of your enterprise. Discover new clients and potential partners, and be discovered in the new world of the Semantic Web!

Further Reading

Something went wrong with that request. Please try again.