Friday, November 08, 2013

#divcon - #Search, #statistics and the proof of the pudding

When you search #Wikidata, chances are that you have a better result for your language than in Wikipedia. As the proof of the pudding is in the eating, there are 5,801,862 items with a label in English and, there are 5,319,467 items with a link to Wikipedia. Based on this comparison, there are 482,395 more labels than links.

These numbers are provided yet again by a new tool from Magnus. This is not to say that we cover the sum of all knowledge; Google does a better job at that. What these statistics show is that we have a perspective on doing a better job at providing information.

When we look at people with the Reasonator, you get a glimpse of what can be done with visualisation. It shows the information in an ordered way and it is easy to notice what is missing. When we present you with all the female painters, it does not take much to notice that women from Scandinavia, East Europe or Asia are mostly missing. This is not even to say that the lady painters from the Anglo-Saxon cultures are well covered in Wikidata. It is however exactly the point; this lists motivates to find more lady painters. You are welcome to fiddle with the query to find data out that motivates you.

These tools are a hack; they demonstrate a future that needs to be build for us all to use. As the future gets build, it needs to scale and it needs to be usable in any language. There are so many pieces to that puzzle, the questions are many; what will be build first and, who is going to do all that and when will we be happy.

No comments: