Time To Experiment With Dbpedia?
Posted by Brian Kelly on 19 November 2009
A the recent CETIS 2009 conference I attended a session on “Universities and Colleges in the Giant Global Graph” facilitated by Adam Cooper. There was a feeling that the initial discussions had perhaps focussed too much on detailed technical aspects about Linked Data, and had failed to address the interests of the senior managers present, who were more interested in what Linked Data could do, rather than whether, for example, RDF should be a mandatory requirement of a Linked Data service.
After the coffee break there was a discussion of ways in which Linked Data could be used in an educational context. One suggestion I made was that as DBpedia (an RDF representation of the content of Wikipedia) provides access to a large amount of Linked Data we should be exploring ways in which we can make use of DBpedia to provide examples of what Linked Data can provide. After all if the data is available shouldn’t we be using it to support advocacy work rather than trying to seek funding to create Linked Data resources?
I was told that DBpedia provides access to structured text boxes in Wikipedia entries, such as the factual entries for Universities (as illustrated).
Could, I wonder, this information be used to demonstrate how such Web pages can be processed as entries in a database rather than just text to be displayed for reading?
So I started experimenting with the DBpedia Faceted Browser.
In the search box I typed “University” and found there were 9,490 entries. After selecting this search option I was then presented with a number of pre-programmed searches such as Country (193 entries for the UK) and City (60 entries for London), I could also search for universities which were established in a particular year (or range).
Searching for universities founded in 1966 I found there were 107 results, including the University of Bath, as shown below.
Can we do more, I wonder, with the RDF data which is already available in DBpedia?
- Can we use this example to demonstrate the importance of data as opposed to a HTML representation of data designed for viewing?
- Can we develop of queries which people may find useful?
- Can we think of data about institutions which could be stored in Wikipedia to allow further queries to be answered?
I also wonder whether it would be possible to go beyond running queries based on the content of the University entries in Wikipedia and explore related pages.
An opportunity for experimentation, perhaps?