Page MenuHomePhabricator

Examine Fataburen metadata and plan the upload
Closed, ResolvedPublic

Description

Examine the Fataburen metadata we got and plan the upload, incl. what cleanup will be necessary and how much relevant information is already on Wikidata.

Event Timeline

So far we've got:

  • list of articles with author, date, title
  • list of authors from the museum's authority database ca 600 rows. They don't have any external auth identifiers, or any other info beyond name

Issues to investigate:

  • References

is a mix'n'match a good approach to find existing wikidata items? 600 should be pretty fast and we could probably get some volunteer help.

Looked some more at the authors. OpenRefine matched 213 with high certainty, and gave suggestions for manual review for another 100. Some of the matches/hints are, I assume, rubbish, as I'm guessing your Marianne Andersson is neither this or this one.

As the dataset contains many 'ordinary' names I would be very careful with automatic matches for those.

Re: Mix n match – to import a dataset you need to give the URL of the catalog as a required parameter. If you just published it on Kulturnav or something it would solve this problem :P

In the end, if no item can be identified, the author name will be added as author name string, which is the usual way to go when dealing with scholarly articles. Then volunteers can help replacing those using https://tools.wmflabs.org/author-disambiguator

I think my next step will be to order the authors by how many articles they wrote and make sure the top ones are matched correctly, as those are likely to be more obvious on Wikidata.