Images, metadata, and gamification of Wikidata

Magnus Manske

��https://tinyurl.com/wdmmsp

E pluribus unum

https://tinyurl.com/wdmmsp

The Wikidata ecosystem

https://tinyurl.com/wdmmsp

Using Wikidata on Wikipedia et al.

• Language links
• Data transclusion via template (e.g. Infobox)
• Relevant Wikidata edits in Recent Changes
• Listeria list generation

https://tinyurl.com/wdmmsp

Example: Commons infobox, made entirely from Wikidata

Category:Monumento Drusa

https://tinyurl.com/wdmmsp

Listeria bot

• Uses SPARQL queries in templates
• Updates lists on all Wikimedia wikis
• Keep track of changes, missing articles on your Wikipedia etc.

https://tinyurl.com/wdmmsp

Wikisource and Wikidata

Scanned pages�on Commons

https://tinyurl.com/wdmmsp

External identifiers

• Wikidata as a linked data hub
• “Don’t store all the data, but link to it”
• Path from external databases, through Wikidata, to other external databases (example query)
• Queryable via SPARQL, even across databases
• External IDs now searchable on Wikidata
• Find items by external ID in Resolver
• Get correlations of two external IDs via BEACON

VIAF

Q15994492

Q30541962

Q18970312

PF3D7_0810800

Simon Iain Hay

​

author of:

Spatiotemporal mathematical modelling of mutations of the dhps gene in African Plasmodium falciparum.

​

Main subject:

​

PPPK-DHPS�(gene in Plasmodium falciparum)

Entry about that gene in GeneDB

Person in VIAF

https://tinyurl.com/wdmmsp

Searching for external IDs

Use SPARQL, e.g. SELECT ?q { ?q wdt:P214 “VIAF_number” }

​

​

Or, just new this week: Search on Wikidata:

https://tinyurl.com/wdmmsp

Mix’n’match

• Takes “catalogs” of external identifiers
• Name, description, type, URL
• Match them to Wikidata
• Semi- and fully automated matches
• Support functions for manual matching
• Can also edit Wikidata for you
• Allows third parties to download matches
• Over 1100 catalogs online

https://tinyurl.com/wdmmsp

OpenRefine

• “Reverse Mix’n’match”
• Start with a CSV file
• Generic, stand-alone application
• Widely used in GLAM
• Now works directly with Wikidata
• Find by (fuzzy) name/type match
• Video tutorials

​

https://tinyurl.com/wdmmsp

Science hub

• Internally connecting genes, proteins, phenotypes, publications, authors, GO terms, locations etc.
• Externally connecting to many scientific databases (“semantic web”)
• Wikidata holds key data, but is not trying to replicate all the databases!
• Scholia tool

https://tinyurl.com/wdmmsp

ContentMine

• Parse scientific literature
• 30.000 pages/day
• Extract keywords and data
• Link up with Wikidata
• Funded by WMF since 2017

https://tinyurl.com/wdmmsp

GLAM (Galleries, Libraries, Archives, Museums)

• Institutions
• Creators
• Artwork
• >300 000 paintings in Wikidata, >284 000 linking to Creator item,�~ 93 000 with Commons file
• Viewer tool: Crotos

https://tinyurl.com/wdmmsp

Wikibase for Commons

• Store file metadata ({{Information}}, licenses, GPS, EXIF…) in wikibase on Commons
• Machine-readable/-writable/-queryable
• Multilingual
• Using Wikidata items where appropriate
• New multi-content revisions under development

https://tinyurl.com/wdmmsp

WikiShootMe

• Show Wikidata items, Wikipedia articles, Commons images, Flickr free images on the same map
• Shows Wikidata items without an image (red circles)
• Upload files and add them to Wikidata (works also from mobile!)

https://tinyurl.com/wdmmsp

PictureThis!

• Simple, mobile friendly tool
• Have eyes on an object/person and wonder if there is an image on Wikidata?�Just start typing the label…
• Can also list objects near you (with detailed address, �where available)
• Directly upload from the tool
• Great for going through your picture collection!

https://tinyurl.com/wdmmsp

Free Image Search Tool

• Find Commons files for Wikidata items with no image, audio etc.
• Can use Wikipedia articles, Commons site search, coordinates in Commons files, Commons categories etc.
• Option to hide unsuitable matches in the future (so results get more relevant over time)

https://tinyurl.com/wdmmsp

File Candidates

• >200K pre-matched file candidates for Wikidata
• Using full-text search
• Free files from Commons & Flickr (and maybe more?)
• One-click Flickr => Commons
• One-/two-click �Commons => Wikidata
• Also as files=>items

https://tinyurl.com/wdmmsp

Wikidata game

• Wikidata lends itself to “atomic edits”
• Many good additions to Wikidata can be automatically extrapolated from existing data, but require human oversight
• Example: Images on a Wikipedia article are likely to depict the article subject, but many false negatives
• Simples Yes/No answers, skip-able
• “Distributed Game” lets others add their own sub-games via JSON
• Designed for mobile

https://tinyurl.com/wdmmsp

TABernacle

• Use a list of items or a SPARQL query
• See values for specific properties in a table
• Add item links right in the table

https://tinyurl.com/wdmmsp

Lexemes

• New entity type, like item or property
• Word forms
• Part of an effort for a free, machine-readable dictionary
• Links between Items and Lexemes
• Relationship with Wiktionary is … under development

https://tinyurl.com/wdmmsp

Fin

Activities

https://tinyurl.com/wdmmsp

Tasks

• Mix’n’match
• Academia Brasileira de Letras
• Academia Brasileira de Ciências
• Brazilian Olympians
• WikiShootMe!
• São Paulo
• Free Image Search Tool
• Items related to São Paulo
• (as above, but by coordinates)
• SPARQL
• Brazilian food (only one item???)

https://tinyurl.com/wdmmsp

• Articles on pt.wikipedia.org without Wikidata item
• ​

https://tinyurl.com/wdmmsp