Wikidata as universal (library) thesaurus

Theo van Veen

Playlists: 'wikidatacon2017' videos starting here / audio / related events

At the research department of KB we linked named entities in newspaper articles to Wikidata resource descriptions offering context information for those entities and enabling searching based on properties obtained from Wikidata. In fact we use Wikidata as a thesaurus for entities in newspaper articles. This functionality can easily be extended to the library catalogue. In bibliographic records, resources get an identifier to distinguish between resources with the same name and for combining name variants. In most cases, this identifier is a URL, also giving access to context information. But how can we find out whether resources mentioned in different catalogues are the same? A lot of links between those identifiers are available in VIAF. So from the identifier in one catalogue we find the VIAF identifier and from the VIAF identifier we find the ISNI identifier and from the ISNI identifier we find the Wikidata identifier. But suppose we were to use the Wikidata identifier directly in our bibliographic records. Then we could provide more functionality without all this "hassle". And when other libraries do the same, it makes searching a specific resource in other catalogues a lot easier. A query like: “give me all books with author X, where X is the same as Y in Wikidata and were Y is the same as Z from the KB thesaurus” is more complicated than “give me all books with Wikidata identifier Y”. In this presentation I will show some functionality available in the KB research environment using Wikidata and I will discuss the use of Wikidata as thesaurus for individual libraries wanting more semantic functionality with less effort. Furthermore I will discuss the problems of conflicting viewpoints, responsibility for the contents and the mixed use of Wikidata and local and other thesauri.