Sunday, January 05, 2014

Linking to #external #sources

Recently references to #Freebase have been added to #Wikidata. When you look at the #statistics, a significant number of references have been added. It the last period 595,778 statements were added that refer to other sources.

When you consider that there are over 14 million items in Wikidata, many more references are to be expected from Freebase. Then again, making the link to other resources is not trivial. As Wikidata gains more statements for its items it will become possible to make those links.

Many of the items that are linked to Freebase, are linked to other sources as well. What I wonder is to what extend these other sources have their own links to other sources. Combining all these links in Wikidata will theoretically make it possible to find information that is currently missing in Wikidata.

Many of these sources have a known error rate. One area in Wikidata were a realistic error rate can be determined is in the interwiki links. The rest is incomplete and immature. We do know that a lot of work needs to be done on the interwiki links; many issues exist that have not been resolved.

Given the lack of quality and quantiy of the existing Wikidata data, importing data from external sources makes sense. Comparing the data with all the external sources makes sense. With more data it becomes easier to infer knowledge from Wikidata, it will become easier to compare with external sources but more important, Wikidata will be better able to provide information to the projects that rely on it.
Thanks,
      GerardM


No comments: