User:Zazpot/Charles Matthews session 2017 06 29

Session: Introduction to Wikidata.

Charles's notes for the session: d:User:Charles_Matthews/Training_supporting_links2.

Wikidata item entries
Each Wikidata item entry has five sections:


 * 1) The Q number (e.g. David Beckham's page gives Q10520).
 * 2) Description with aliases, in natural languages.
 * 3) Statements of substantive kinds.
 * 4) Statements relating the item to identifiers in other databases.
 * 5) Interwiki links to corresponding articles on other WMF wikis.

Improved categorisation
Reduction of gender bias. Currently, articles about French women engineers - to take an example - are not adequately tagged with that category on the English Wikipedia, due to lack of maintenance, i.e. lack of editor attention. I found 35 French female engineers using Wikidata. The corresponding Wikipedia category misses 13 of them. So, this gender bias in the English Wikipedia is remediable via Wikidata.

Highlight missing images
Academic and other institutions, that may be the sole holders of images of prominent people, are not always amenable to licensing those images to the commons.

Wikidata can help us convince those institutions to freely license such images, by letting us establish concretely just which people are missing images, and comparing their coverage to other prominent people in the same field. Not only can this be done via lists

Mapping

 * Cathedrals (shows pattern that they tend to appear along major roads, esp. Roman roads)
 * Libraries (note the importance of querying subcategories)
 * Photo-mapping: WikiShootMe, etc.

Caveats
The Independent's use of Wikidata to analyse the Panama Papers shows evidence of confirmation(?) bias. The Independent suggests that the largest category of people mentioned is politicians. But actually, that might just show that politicians are better categorised in Wikidata than people in most other professions.

Hands-on session
Some great SPARQL query examples. Introduced WDQ.

Noted importance to be aware of entries about fictional characters, etc. (E.g. searching for Presidents of the USA brings up fictional characters like David Palmer.)