Skip to 0 minutes and 5 seconds BARRY NORTON: In this course, one example that we’ve given of applications built using Linked Data is in the music domain. So MusicBrainz, in itself, is a useful resource which we could look at as a graph representing RDF and query of SPARQL. But what becomes really interesting is integrating information from other knowledge graphs. So we give a flavour of how Wikipedia information, which is stronger on the biography of people for instance, can be brought together with MusicBrainz using RDF. And we can form applications that talk both about the music output and about the lives of music artists. Another interesting application of Linked Data technology can be found in the cultural heritage domain.
Skip to 0 minutes and 47 seconds Here we have museums publishing collections information as linked data. We also have third parties who represent expert knowledge that can be used to bring together that information. So a prominent example recently was the release of linked data sets by the Getty Institute. They describe types of artefacts, materials, and techniques, and we can use those expert understandings to bring together data sets from different institutions. Many applications will have a component of interesting geographic information. We might want to bring together a business that covers many geographic territories or products and services from those areas.
Skip to 1 minute and 34 seconds Not every business wants to build a data set about the geography in itself, especially not one open to new applications that weren’t foretold at the time that a database was convened. But by linking to data sets that come from experts in a area, for instance, here in Southampton, the Ordnance Survey, we can take advantage of various levels of description of geography. So immediately we might just be interested in places but then we might link to post codes in particular applications that we then built on top of our data set. There are then other data sets related to the geography that we might bring into whole new applications built on top of our data set.
Skip to 2 minutes and 16 seconds So for instance there’s logistic and transport information that we might bring in from governmental or private sources to build whole new applications. In particular, by representing our knowledge in a graph of linked data, we then have access to semantic web technologies on a reasoning that allow us to take those facts that we’ve explicitly represented to infer what implicitly is meant by those and derive new knowledge. This is often a good technique in integrating data from different sources. And different people may have found certain characteristics of primary importance but also infer common meaning that can be used to bring together their knowledge.
Skip to 2 minutes and 58 seconds Other related technologies that learners might be interested in beyond this course are those that allow the production of linked data from non-graph sources. So for instance, how can we take natural language text that a human can read easily and draw out the knowledge in a graph-like way to integrate with our systems. Another interesting challenge, where approaches exist within the Linked Data technology stack, is taking sources that are not themselves graphs but drawing out the knowledge and representing that as a graph. So we might want, for instance, to take a narrative text that a human could read but draw out the important parts of knowledge and represent those in a graph.
Skip to 3 minutes and 39 seconds We might also take structured sources, for instance, spreadsheets and relational databases and draw out the pertinent features into a graph.
Developing real world applications
Watch Dr Barry Norton describing some real world applications that have Linked Data as their underlying technology.
Barry mentions resources such as MusicBrainz, which you will discover in more detail later on (step 2.2.). MusicBrainz uses Linked Data to collect music metadata from different sources, such as Wikipedia, and mashes it up.
Another example Barry mentions is that of the datasets released in Linked Data formats by the Getty Institute. Such linked datasets are available in the form of vocabularies such as the Cultural Objects Name Authority (CONA)
Why not to try and make a search in these applications? You could search for a song or artist in MusicBrainz, or for an artifact in one of the Getty vocabularies. What is it that you find different from a conventional search engine?
To explore the possibilities of linked data browsers and mashups (which combine data from many sources), look at these examples of working websites based on semantic web technology.
The Marbles site, created at the Berlin Freie Universitat, allows you to view presentations based on RDF data from multiple sources that are distinguished visually by marbles of different colours. At the bottom of the page, these colours are indexed to the URIs of the sources. The value of such presentations is that they show at a glance how strongly the information is attested among the various datasets, thus providing some indication of its reliability. Applications like Marbles that exploit multiple datasets are sometimes called mash-ups.
DBpedia Mobile is an application for mobile devices (phones, pads) which uses location detection in order to offer information from DBpedia on the user’s current neighbourhood. The user sees a map the neighbourhood on which various features of potential interest are labelled; clicking on a label opens a pane giving information generated from DPpedia and linked datasets. The application uses the Marbles Linked Data Browser (see above).
Some other sites
For further examples of sites using Linked Data, see the following.
The BBC has launched a music portal based on Linked Data.
The University of Leipzig has a community project providing street map information based on Linked Data.
In 2009 the US and UK governments made commitments to open data. This portal contains the US datasets.
Available with over 8000 datasets published as of 2013.
© University of Southampton 2016