-
Remember life before Wikipedia?
-
"An encyclopedia that anyone can edit"
-
knowledge base
-
it's a wiki
-
community model
-
link Wikipedia language editions
-
reuse statements across Wikipedia projects
-
provide complex query capabilities
It's awesome, especially if you're into data!
Factual claims are stored as statements
subject -- predicate -- object
thing -- relationship -- thing
item -- property -- value
Similar to RDF (and mapped to its model)
Independent of language (identifiers vs. names)
-
entities, labels, descriptions, statements
-
types of entities
- items (have wiki-links)
- properties (have data types and constraint statements)
- lexemes
Find item of your home town, school...
-
Statement details
- properties
- qualifiers
- references
http://bit.ly/wikidata-onepage
https://www.wikidata.org/wiki/Wikidata:In_one_page
-
A hub in the linked open data web
-
Wikidata properties for identifiers
-
One possible overview:
https://www.wikidata.org/wiki/Wikidata:Identifiers
In groups of 2-3:
-
add/extend Wikidata items on some of your professors
-
see existing professor items as boilerplates
-
collect questions for afterwards
-
more than 55 million media files
-
not as shiny as Instagram, YouTube, Flickr...
-
but Open Content, no commercial interest!
-
community model
-
quite "unstructured"
Migration of Wikimedia Commons to Wikibase (2017-2019)
-
every media file is an entity
- multilingual media file captions
- statements about media files
-
properties reused from Wikidata, e.g. depicts (P180)
-
work in progress (e.g. no SPARQL yet)
More information at https://commons.wikimedia.org/wiki/Commons:Structured_data
Intoduction of three new types of entities in 2018:
- Lexemes (L)
- Forms (F)
- Senses (S)
https://www.wikidata.org/wiki/Wikidata:Lexicographical_data
Sample application: http://auregann.fr/derdiedas/
-
entities, labels, descriptions, statements
- items (have wiki-links)
- properties (have data types and constraint statements)
- lexemes
-
statements
- property
- qualifiers
- references
-
Wikidata query service (SPARQL)
-
Several tools and programming libraries
-
(big) data dumps
-
Coverage is very inconsistent
-
Data modeling is instable
-
Qualifiers and references help to improve quality
- but not used as much
- harder to query
-
Working with Wikidata is like doing data science:
cleaning data & fighting with software
-
People are not paid
-
Nobody has a full overview
-
Tools (plenty!) come and go
-
Be nice and allow misunderstandings