lördag 4 november 2017

Wikidata checks gender diversity

Question: Why is there a big difference between Sweden and Finland in the number women documented at SBL vs. Kansallisbiografia?

Below external identifiers in Wikidata and gender diversity

Feels like Wikidata can be a way to "benchmark" gender diversity between external source e.g. 

P3222 Nationalencyklopedin Online - Ratio: 11%

Male: 11724 - Female: 1576


P3217 Swedish National Archive SBL - Ratio: 6 %

Male: 6901 - Female: 508

P2538 Nationalmuseum Sweden artist - Ratio: 10 %

Male: 4088 - Female: 460

P2180 Finnish Kansallisbiografia - Ratio: 14,5 %

Male: 4270 - Female: 735

P4357 Swedish Musikverket - Ratio: 50 %

Male: 792 - Female: 801

P3360 Nobel Prize People Nomination - Ratio: 5 %

Male: 3102 - Female: 147

P8024 Nobel Prize ID  - Ratio: 5.5 %

Male: 866 - Female: 53

P1138 Kunstindeks Danmark Artist  - Ratio: 10 %

Male: 4177 - Female: 468

P906 Libris - Ratio: 15 %

Male: 50266 - Female: 8957

Next step ?

I can see as Wikidata gets more information we can start measure
  1. How well museums document minorites
  2. Minority languages
  3. People working in the whole country and not just people in e.g. the capital of Stockholm
  4. Start create maps like below but for how well they support diversity

1 kommentar:

  1. The reason you see a higher percentage of males in older biographical datasets is often also in relation to the size of the matched data. So in this case, the Swedish dataset has way fewer matches than the Finnish dataset. In my work on Dutch artists and matches in Wikidata, I have noticed that over time, the percentage female creeps up from around 3% to as high as 11%. This is because the inherent systemic bias of Western society measures the "kings and conquerors" first and "queens and mothers" later. Generally the women are included only as linking pins to important sons-in-law, etc.