Wikidata:WikiProject Authority control

This WikiProject aims to the improvement of the quantity and quality of external identifiers present on Wikidata.

If you have ideas, please open a new thread in the talk of the WikiProject.

Overview on external identifiersEdit

Most useful tools and gadgetsEdit

Main tools for coreferencing:

Main gadgets for coreferencing:

  • User:Magnus Manske/mixnmatch gadget.js: when you open an item, you see all the Mix'n'match entries automatically matched to the item and you can manually confirm the correct automatches
  • User:Bargioni/MnM ext2.js: little extension of the previous gadget, which allows to easily remove the incorrect automatches (one click on X) and to eventually mark entries as not applicable to Wikidata (two clicks on X)

VIAFEdit

For information about the relationship between Wikidata and Virtual International Authority File (Q54919), see Wikidata:VIAF and its subpages

AATEdit

The Art & Architecture Thesaurus (Q611299) from the Getty Research Institute (Q11203476) is a crucial multilingual thesaurus in cultural heritage, with 53,249 concepts as of 11 September 2020. See http://vocab.getty.edu/sparql

    select (count(*) as ?c) {
      ?x a skos:Concept; skos:inScheme aat: }

Of these, 22,000 (54.4%) are mapped to 21,773 Wikidata items, also as of 11 September 2020. (Live query.)

AAT is actively coreferenced on Mix-n-Match.

For historical information about the relationship between Wikidata and AAT, see the archived material at Coreference AAT.

ULANEdit

The Union List of Artist Names (Q2494649), also from the Getty Research Institute (Q11203476), is a dataset of entities in the art world, primarily artists but also museums, galleries, organizations, and companies, with 312,079 entries as of 12 September 2020. See http://vocab.getty.edu/sparql

    select (count(*) as ?c) {
      ?x a skos:Concept; skos:inScheme ulan: }

Of these, 79,529 (43.2%) are mapped to 88,415 Wikidata items, also as of 12 September 2020. (Live query.) 45,032 are preliminarily matched based on labels (names) and need to be verified (expect a high percentage of false positives in this group).

ULAN is actively coreferenced on Mix-n-Match, but this dataset requires some manual review after import. Items to watch for:

  • Mix'n'Match contains only 183,912 of the ULAN items (those representing humans).
  • A few ULAN names are formatted LAST NAME, FIRST NAME, and will be imported that way by Mix'n'Match.
  • Some punctuation in names will be imported with escape characters (//) by Mix'n'Match; these need to be removed.
  • Mix'n'Match may import floruit (Q36424) or "active" dates as birth and death dates; these should be deprecated, ideally with <reason for deprecation> work period dates imported or interpreted as birth/death dates (Q80833195). The active dates can be correctly added using floruit (P1317) or work period (start) (P2031) and work period (end) (P2032).
  • ULAN is coreferenced in VIAF.
  • ULAN contains values for sex or gender (P21) and occupation (P106), but these are not imported by Mix'n'Match.
  • ULAN contains many alternative names and spellings, which are not captured by Mix'n'Match but can be very helpful for coreferencing to other sources. Adding these as aliases by hand is good!

RKD artistsEdit

RKDartists (Q17299517) is a database of artist biographies from the Netherlands Institute for Art History (Q758610). The database is bilingual (Dutch and English). Of ~422K entries in the database, about 40% are redirects to other items and have been marked at "not applicable" to Wikidata. 92,346 entries are mapped to 88,415 Wikidata items (as of 15 September 2020). 55,389 entries are preliminarily matched based on labels (names) and need to be verified (expect a high percentage of false positives in this group).

RKDartists is actively coreferenced on Mix-n-Match. The structured data in RKDartists is very robust, and once an RKDartist ID has been mapped to a human in Wikidata, a Bot will automatically create statements for the available structured data, with references. The statements added by the Bot include labels in some European languages, date and place of birth, date and place of death, occupation, floruit or work period start/end dates, and work locations with start/end times.

Note:

  • Items created from RKDartists ID using Mix'n'Match may contain EN descriptions in Dutch; these should be replaced.
  • The database contains alternate forms and spellings of names, but these are not automatically added as aliases. Adding them manually will help in coreferencing to other datasets such as ULAN.

HistoryEdit

Please add here references, blogs etc on the topic. For news prior to 2019 see the archive.

Tweet using tag #coreferencing.

  • 2020: ...

Useful resourcesEdit

ParticipantsEdit

Please become members of this project!


[+ Add yourself to the list]

The participants listed below can be notified using the following template in discussions:

{{Ping project|Authority control}}

Related WikiProjectsEdit