Wikidata

Wikidata is a collaboratively edited knowledge base hosted by the Wikimedia Foundation.

The content is maintained by its users community.

The database is published under CC0 (public domain). Other content as CC BY-SA. MediaWiki software as GNU GPL.

Technique

Wikidata is a triplestore database, using the linked open data (LoD/RDF) and URI/PID techniques to interconnect with other databases. It is implemented as a Wikibase. It contains recursive statements (premises) of subject-property-object/value tuples to describe an item, identified by a Qid. The physical database contains one single table with 3 columns. The logical data model is implicitly defined on top of the relations amongst the items.

Every object is uniquely defined by a label in any language, a description, and an instance or a subclass. Objects can also have aliases. Statements are augmented with qualifiers, and references.

The physical structure of the database almost never changes. Instead, the logical datamode changes by defining new properties. Properties rarely expire (e.g. P642). Properties have a data type.

Unlike a relational database, it does not have primary nor secondary indexes. It doesn't implement hard constraints. Nor it is ACID transactional. The system tries to be eventual consistent, allowing to describe a maximum of information.

Purpose

Wikidata has the goal of collecting, structuring, and distributing free knowledge, just like Wikipedia, and the other Wikimedia platforms. For Wikipedia this happens in the form of encyclopedic texts, in Wikidata it is structured and linked data. Wikimedia Commons stores media files, and has a Wikibase instance of Structured data on Commons (SDoC). The 3 platforms are interconnected. They can all be updated via Pywikibot.

History and background

Before Wikidata existed, the Wikipedia sidebar section In other languages held links to articles in other languages. They were registered manually in each single article. That eventually resulted in numerous interwiki inconsistencies and errors, and the more a subject was described in multiple languages, it became more complex to maintain. Think about half a matrix of 300 languages resulting in 45000 unique potential combinations. For each article title change, potentially more than 300 Wikpedia language pages would need to be updated.

On 29 October 2012 Wikidata went live. The wiki was empty, and volunteers first started to import all the sitelinks from Wikipedia, removing the interwiki links from the Wikipedia pages. Q1 "universe", Q2 "earth", and Q3 "life", were the first items. The sitelinks with other Wikimedia platforms were added later.

As of 2013 Wikidata became a generic triplestore database, to document the world. The content can be reused on Wikipedia via infoboxes and other tools.

In 2017 Wikidata Query was added as a query tool to create reports.

In 2018 the Structured Data on Commons (SDC/SDoC) Wikibase instance was implemented. This allowed creating sitelinks from Wikimedia Commons to Wikidata, using the same properties as in Wikidata.

Functionality

Data model

The physical data model of Wikidata is 1 huge table and 3 columns: subject item - property - object item/value. The statements based on predicates form a triplestore or RDF database by which the whole world can be described. This can power artificial intelligence.

Useful gadgets

You can set the following via the user preferences option (by default disabled)

[ ] Merge: This script adds a tool for merging items
[ ] moveClaim: A tool to move or copy a statement from one entity to another
[ ] labelLister: Show and edit labels, descriptions and aliases from other languages
[ ] Reasonator: Adds a link to Reasonator on every item
[ ] relateditems: Adds a button to the bottom of item pages to display inverse statements
[ ] EasyQuery: A tool to add an icon for finding items with same statements easier

Contacts

Geertivp

Wikimedia Belgium has no authority in Wikidata. It is the community of volunteers on Wikidata who makes the decisions about content. The physical website infrastructure is owned and made available by the Wikimedia Foundation. Wikimedia Belgium is founded to support local volunteers and organisations in Belgium in working with Wikidata and the other Wikimedia platforms. We have years of experience in helping organisations with using Wikidata and adding data from collections to Wikidata.

Timeline

30 October 2012 - Start of Wikidata with first content
16 December 2014 - Freebase announced to retire and to move their data to Wikidata (FAQ) (archived)
26 March 2015 - VIAF is moving to Wikidata (archived)
28-29 October 2017 - WikidataCon in Berlin
Wikimedia Hackathon 2019 in Prague
Structured data on Commons

Projects in Belgium

Below is an incomplete list of projects done in Belgium on Wikidata.

October 2015-January 2016 - Import data from Flemish art collections to Wikidata (info)
March 2017- 2018 - Import data from the Kunstenpunt database to Wikidata (info)
September 2017-November 2017 - Import of the monuments of the Brussels Capital Region to Wikidata (info)
November 2017-May 2018 - Import of archive locations from Archiefbank Vlaanderen to Wikidata (info)
September-November 2018 - Import of the monuments of the German-speaking Community to Wikidata (info)
29 October 2018 - Wikidata workshop organised in the Royal Library of Belgium in Brussels (info | event page)
29 October 2018 - Wikidata 6th birthday party in the Royal Library of Belgium in Brussels (info | event page)
3-5 June 2019 - d:Wikidata:Events/UGent Wikidata and Wikibase Workshop 2019 (photos)

Known problems

This result is cached and can be outdated for up to 59 minutes: Refresh the cache via the "purge" option

Instructional videos

YouTube

Hoe Wikidata items aanmaken en bijwerken
Wikimedia koppelen met OpenStreetMap