1 of 33

@EvoMRI @readermeter

Daniel Mietchen Dario Taraborelli

Wikidata and Wikibase as global platforms for democratizing data publishing

SciDataCon 2018 • Gaborone, 6 November 2018

University of Virginia Wikimedia Foundation

Abstract: Over the past several years, interest in leveraging Wikidata as an open knowledge base has been growing in areas ranging from cultural heritage institutions and libraries to research and technology organizations. With its inherent multilinguality, human-editable interface, community-driven approach to data modeling and curation, systemic connection to Wikipedia and sister projects and alignment with the FAIR Principles for sharing data, there is an emerging consensus that Wikidata represents a significant step for turning Linked Open Data into a practical and useful technology. It provides a bridge between the many siloed knowledge bases that have emerged in the more than two decades since Tim Berners-Lee first proposed the notion of the semantic web. Despite this growing interest, there is still a fundamental lack of understanding of what Wikidata's strengths and weaknesses are, or those of Wikibase —the engine behind Wikidata. While experiments have been proliferating with the use of Wikibase, it is still very much an open question how to create a sustainable, federated ecosystem of knowledge bases to support open research, and what the social, technical, institutional barriers are towards this vision . We would like to use this session as an opportunity to identify and discuss user stories from researchers and practitioners working on open research infrastructure and understand what a model could be to get more groups onboard to design a possible path towards this vision.

2 of 33

Two main avenues for

democratizing data publishing

A knowledge graph that anyone can edit and query in their own language

A Wikidata-compatible graph database that anyone can set up and federate

3 of 33

Wilkinson et al. (2016) doi.org/10.1038/sdata.2016.18 [image: fosteropenscience.eu CC0]

FAIR data platforms

4 of 33

5 of 33

Wikidata is to data what Wikipedia is to text

All data CC0
Anybody can contribute
Covers all domains of knowledge
Fully version controlled and collaborative
Integrated with the semantic web via RDF / open APIs
High performance query engine (SPARQL)
Stable. Not tied to short-term funding cycles
Actively developed, full stack is open source
Active community. Fastest growing Wikimedia project

6 of 33

�550M statements • 760M edits

[as of October 2018]

50 million entities

7 of 33

Types of content in Wikidata

People
Places
Taxa
Buildings
Organisations
Artworks
Events
Astronomical Bodies

...

Chemicals
Processes
Theorems
Concepts
Creative works
Journals
Publishers
Meta-items

...

8 of 33

Wikidata for Research

https://www.wikidata.org/wiki/Wikidata:WikiProject_Wikidata_for_research

9 of 33

Data provenance of a Wikidata statements by outlet, publisher and funder

Zika virus • Q202864

TAXON�

has natural reservoir • P1605

Aedes hensilli • Q14573674�TAXON�

stated in • P248

Aedes hensilli as a potential vector of Chikungunya and Zika viruses • Q22330738�SCIENTIFIC ARTICLE�

funded by • P859

Centers for Disease Control and Prevention • Q583725�GOVERNMENT AGENCY

published in • P1433

PLOS Neglected Tropical Diseases • Q3359737�SCIENTIFIC JOURNAL�

publisher • P123

Public Library of Science • Q233358�PUBLISHER

10 of 33

Sample of current biomedical content in Wikidata

All human, mouse genes and proteins (swissprot)
All Gene Ontology terms
All Human Disease Ontology terms
All FDA approved drugs
109 reference microbial genomes

Mitraka et al (2015) Semantic Web Applications for the Life Sciences

Burgstaller-Muelbacher et al (2016) Database

Putman et al (2016) Database

11 of 33

Biologists with Canadian citizenship

[SPARQL query]

12 of 33

Institutions where Canadians got their PhD

[SPARQL query]

13 of 33

Co-author graph of McGill-affiliated authors

[SPARQL query]

14 of 33

Award recipients affiliated with McGill

[SPARQL query]

15 of 33

Federation

16 of 33

Wikidata’s identifier mappings

17 of 33

From Wikidata to Wikibase

18 of 33

Why Wikibase?

Linked Jazz

“Started thinking about how our data could live in Wikidata and started investigating feasibility of that possibility.

But we have very esoteric project data that doesn’t seem appropriate to be in Wikidata so begain looking at our own Wikibase instance.”

Matt Miller (2018) Linked Jazz and Wikibase

19 of 33

What’s Wikibase

Wikibase Repository - MediaWiki extension for structured, non-relational data in a central, collaboratively managed repository.

“writing RDF”
Revision control
FAIR by default

Wikibase Client - MediaWiki extension for retrieving and embedding structured data from a central repository into a client wiki.
Query Service that allows to query the contents of a Wikibase installation using SPARQL
A set of reusable components that provide a foundation for tasks in the same domain.

http://wikiba.se/

20 of 33

Data formats in Wikibase

(versus wikitext)

21 of 33

Many faces of Wikibase

22 of 33

Let's take a tour

23 of 33

Properties on the Rhizome Wikibase

24 of 33

A French recording of the word “Canada”

From Lingua Libre

25 of 33

Letters sent by Illuminati

From FactGrid’s SPARQL endpoint

Colour indicates author

26 of 33

Timeline of software repositories

[SPARQL query] on Wikidata

27 of 33

Timeline of Wikibase instances

[SPARQL query] on the Wikibase registry

28 of 33

Wikibase and software repositories

Combined [SPARQL query]

across Wikidata and the Wikibase registry

29 of 33

Further notes

Federation is possible to and from any SPARQL endpoint, not just Wikibase ones (and works fine on mobile)
A Wikibase instance also has a Mediawiki API
Docker container available to install Wikibase
Wikimedia Commons is moving to Wikibase
Ecosystem of Wikidata tools, some being adapted to generic Wikibase instances
Work on using Shape Expressions to share data models across instances
Coordination through series of workshops
Various non-public tests, e.g. at OCLC

30 of 33

Stacy Allison-Cassin (2018) Wikibase & Indigenous Knowledge in the Canadian Context

31 of 33

Wikidata or Wikibase(s)?

Wikidata community	Governance	depends
generic	Granularity	depends
CC0	Licensing	depends
stable	Funding	depends
stable	ID mappings	depends
many	Language(s)	depends

32 of 33

Another approach to democratization of data curation:

citizen science happening on Wikimedia projects

see SciDataCon poster 150

“A wiki approach to collecting, curating and managing citizen science data”

33 of 33

Thank you

growth by Fabio Rinaldi [CC BY], research by Minnie Pigeon [CC BY], �graph by Icon Lauk [CC BY] from the Noun Project

�

Slides mashed up with contributions by �Andy Mabbett and Andra Waagmeester

These slides are adapted from

D. Mietchen, D.Taraborelli (2018) Wikidata, Wikibase, and a federated ecosystem of structured knowledge for open science. FORCE 2018�doi.org/10.6084/m9.figshare.7195358 [CC BY]