Wikidata:WikiProject LD4 Wikidata Affinity Group/Affinity Group Calls/Meeting Notes/2020-10-06

From Wikidata
Jump to navigation Jump to search

Call details[edit]

Time Stamps[edit]

00:00 Introduction and announcements

03:38 Introduction to Wikidata and Entity Explosion

10:39 Outline - Entity Explosion explained

12:47 Wikidata: The Rosetta Stone of the Internet

21:37 Using Entity Explosion

37:32 Multi-language live syllabus

39:30 A workflow for contributing and integrating your data sets

44:46 Reap the rewards of your integrated data

47:28 Final thoughts

52:04 Questions and discussion

Presentation material[edit]

Notes[edit]

  • Entity Explosion is a browser extension to discover links and information about the same topic on other sites
  • Github site for code changes
  • Linked data allows us to create complex queries that would normally take extensive research to determine
  • Most properties in Wikidata are links to external identifiers (not links to other Wikidata items)
  • Can use external identifiers to go to other sites or cross-reference two lists and convert from one to another
  • Wikidata has an incompleteness bias (although it’s improving) that other databases can help mitigate
  • In properties, “formatter URL” takes an ID and turns it into a URL
    • There are also mobile formatter URL, third-party formatter URL, and format as regular expression properties
  • Entity Explosion installs in your browser, and if it sees one of those formats in a URL, thinks it could be linked to an item in Wikidata
    • If it is the Entity Explosion icon changes from gray to red, and if you click the red link EE sends a query to Wikidata
    • EE sends the entire string to a the Wikidata Query Service to check what item the URL corresponds to
    • Then if it is found it turns green, locates the Q ID, and can tell us what the Wikidata item it is, and anything it wants about that item (including links to other sites)
    • You don’t need to be a Wikidata editor to use  this
  • May persuade users to add data to Wikidata if data is missing
  • Shows translation of properties, translating properties is a great service
  • EE is meant to work on every Wikimedia site
  • Hoping to match up “described at URL” and home pages, but not there yet
  • Librarian question: Who is the most famous child of a librarian?
  • Use cases for academia, hobbies, etc. (Example: Click and see who a person is on Twitter, what their qualifications are)
  • Suggests putting it in extensions, and see when it lights up
  • DOIs often redirect to publisher pages, which presents a challenge
  • Free, open source, every language, privacy protected (only sends current site and only when clicking), data is live from Wikidata each time
  • Written in Javascript calling SPARQL
  • Does do a couple of broad queries on (re)install
  • If a site reorganizes itself, the actual identifier might be preserved and the only update needed is to change the format on the Wikidata property page. If a site totally changes their identifiers, would need to create a new Wikidata property for the new identifier
  • Reviews add to visibility in app stores
  • Discussion of translation of syllabi project
  • Consider linking your most interesting data with Wikidata
  • Look for: 1. A different web page for each item including a unique, verifiable ID; make a wikidata property; prepare a list (spreadsheet or scrape from web)
    • 2. Matching with existing items, Mix’n’Match, OpenRefine, etc.
    • 3. Bulk upload of new items
    • 4. Quality checks
  • Getting involved: Download EE, Build Wikidata (propose ID properties, translate ID properties), Help get the word out (share on social media, nominate, mention to colleagues, review and rate)
  • Reap the rewards of your integrated data (labels in every language, corresponding IDs in other languages, error detection by comparison, plethora of relevant information, links to richer textual resources, structured item links/ relationships within dataset, queries to answer broad questions)

Questions[edit]

  • It appears notes were not taken for the Q&A portion. Interested readers can add them from them from the recording!