Meeting 2022 02 24 - wmo-im/wikimedia-pilot GitHub Wiki

Date/Time

Thursday 24 February 2022 13h UTC

Location

MS Teams

Attendees

ECCC / Wikimedia

  • Miguel Trembley
  • Pierre Choffet

TT-WISMD

  • Tom Kralidis

TT-WIGOSMD

  • Rainer Marz
  • Gao Chen

Secretariat

  • Anna Milan
  • Xiaoxia Chen

Agenda

  1. Previous meeting minutes and actions (all)
  2. Implementation update and demo (Miguel/Pierre)
  3. Status/next steps (all)

Notes

  • The team team reviewed actions from last meeting
  • Pierre presented wikimedia-pilot script in the command line. The script will query OSCAR/Surface for list of stations, results go into a local cache and then are transformed to wiki data like structure.
  • Miquel demonstrated the upload of the wikidata for a station in Zambia to wikidata.org. Now the upload is on a per station basis, but in the future, all will be uploaded at once. Only data that has changed from the last upload will be imported. The station identifier links back to record in OSCAR/Surface. See result on wikidata: https://www.wikidata.org/wiki/Q111009324.
    • (RM) Will it be possible to forbid editing of these records? (MT) Miquel watches the pages and will get an email when anything changes on the page. Historical management of 9K+ stations for Canada there have been there was one change that needed intervention. The other changes added value to the records. It's not possible to make not editable. (TK) how long will it take for the script to run on all of OSCAR (PC) days to a week for initial upload. There is additional latency designed into the code (sleeping) so as to not overload wikidata site.
    • (TK) Can we put the source code on the GitHub repo? (PC) the source is currently for demo only, but future code will be in the dev branch
  • (TK) where do we go from here? We have enough for a couple paragraphs in an article in the WMO newsletter (MT) it would be useful to display a subset of stations on a map based on query that is taking advantage of other information in wikidata. (TK) step 1: move code to GitHub. Step 2: upload all stations to wikidata; Step 3: publish article in [newsletter] (Xiaoxia) there will be a newsletter release every two months newsletter subscription; 2022-No.1; 2021-No.1. (MT) we should figure out the license issue first or use only a subset of information that are not considered intellectual property, eg. station name, coordinates, country, id... (PC) open street map handles this issue by including an email/statement from the source institution/db stating that the data can be published to the map.

Actions

  • Anna: email chair of ET-WT to understand options with OSCAR/Surface license
  • Xiaoxia: send link to the newsletter and one can subscribe to future.
  • Tom: summarize decisions from today and send out email

Next meeting

2022-Apr-21 13UTC