Managing Wikimedia Projects through Wikidata
Using Property P6104
What are we learning here?
to be able to:
Prerequisites
The easy part (what we already know ?):
Willingness to learn some technical stuff
Agenda for today
Learning Clinic plan
Methodology developed by Wikimedia Belgium
… during the COVID-19 period (2020)
… requiring to fallback to virtual writing sessions
This way of working still pays off to prepare and organise effective onsite sessions…
Interactions between major platforms
WikiProject flow and timeline
Partner(s)�Project leader(s)�Subject�Budget
Meeting place�Infrastructure�Catering�Security, Friendly spaces
Idea�Concept�Initiative�Requirements
Inventory building�(list of items/pages)�Categories�Excel, OpenRefine, Wikidata
Invitations�Participant registration�Account creation�Number of participants
Activity (physical, virtual)�Presentation, demo�Follow-up�Reporting
Prototype flowchart
Designed by Geertivp.
Presented at the EU GLAM meeting in Prague, 15 September 2022.
Wiki Project management overview
Yearly plan
Project plans
Partners
Strategy
Budget
Volunteers
Experts
Employees
Project setup
Contact a partner
GLAM - Education
Choose a topic
Choose a venue
Multiple sessions?
Gather items
Gather attributes
In Excel or Google Spreadsheet
Project execution
Have the writing session
Make documentation available
References
Write down your experiences
Document the results
(Yearly) metrics
Number of users
Number of articles
Number of photos
Number of sessions
User retention
Repeated sessions
Photo reuse
Project evaluation
What did work?
What could be better?
Spread the news
Newsletters
Page view statistics
Wiki Project setup
Data gathering
Contact a partner
GLAM?
Education?
Choose topic
Choose a venue
Multiple sessions?
Gather items
Gather attributes
In Excel or Google Spreadsheet
OpenRefine
Reconcile
Map to Properties
Amend items
Create items
Link with “Maintained by Wiki Project”
This is the cornerstone of the methodology…
Wikidata
Create the Wiki project
Amend the items manually
Link to the Wikipedia project page
ListeriaBot
Create a subpage
(to not copy static content)
Can have multiple sublists (subproject or query filters)
Wikidata Query
Obtain the list of items
Separate lists?
Or with a filter (by type of instance)
You might use subtables
Wiki Project execution
Program and Events
(Outreach DB)
Create a campaign
Create a program
Create an event
Include some documentation
e.g. link to the project page
Wikipedia project page
Describe the session
Document the venue
Blue subscribe button
List of topics (ListeriaBot subpage)
Session
User register in Outreach dashboard
Users create a user account
Users request a user account
(avoid the proxy server blacklist)
Have the writing session
Statistics
Gather project statistics
Publish the results
Public relations
(local) Newspapers
Project report
Magazines
Newsletters
Meta
Follow-up
Create user welcome pages (script)
(minimum 1 article)
Give feedback to the participants
Publish the draft pages
Monitor deletion by moderators
Evaluate the project
So we need a project page?
Example: Wiki Women Design
What is a WikiProject?
Any collaborative project run by volunteers (Wikipedians, Wikimedians) using the Wikimedia platforms.
Goal: gathering, building, structuring, making available free accessible knowledge.
Typically set up by a country (Chapter) or a (thematic) user group.
Often in collaboration with a library, archive, museum, cultural or heritage group (GLAM).
Requires a venue (class room, Wi-Fi, internet access, projector, possibly food and drinks).
Requires a (small/larger) budget. Can be a repeated session.
Inviting Wikipedians to write on Wikipedia and/or take photographs + coaching.
Supported by using Wikipedia, Wikidata, Wikimedia Commons, and Wikimedia related tools.
Activity types
Usage of main tools in time order (summary)
Excel
OpenRefine
Data gathering:
Data modelling
How to structure?
Data validation
Prepare the upload
Data scheme
Wikidata
Data Upload
Automated via OpenRefine?
Amend the items manually if needed
Any other useful data?
Linked Open Data
Registering existing images (P18)
Wikidata Query
ListeriaBot
Write your query
Execute your query
Get list of items
Map the properties
Validate, verify
Decide on the model:
Create a project table
Wikipedia
Wikimedia
Have the session
User training
Monitor user activity
Coach the users
Amend articles
Rename mature articles to the main namespace
Sitelinks to Wikidata
Events Dashboard
Create a campaign
Create a program
Create sessions
Let users subscribe
Account creation
User assignments
Used platforms
(in time sequence)
�Excel or Google Spreadsheets: compile a list of subjects (data gathering)�Wikidata: create WikiProject; data store and GUI to update items�OpenRefine: reconcile and register the subjects into Wikidata�QuickStatements: to load data into Wikidata (optional)�Pywikibot: automatically update Wikidata (optional)�Wikidata Query: get simple lists of items�ListeriaBot: generate a table of subjects with attributes (red and blue links)�Wikipedia: to create the project page�Programs and Events dashboard: user participation and monitoring�(amend) Wikipedia�Wikimedia Commons + Photo campaigns�Programs and Events dashboard (reporting)�Cloud services (optional)�Excel or Google Spreadsheet
The concept is extremely simple
Spawning multiple possibilities…
P6104: Maintained by WikiProject
Very simple concept…
But the implementation might be more difficult…
You need to define your own ad hoc data model first.
Nothing physical in the database; it is only in your mind...
Wikidata statement (example)
Object or value
Subject
Property
From Online document to Linked open data
5-star diagram from Tim Berners-Lee
Triplestore?
Wikidata as a triplestore:
One huge single table with 3 columns
No proactive data model => ad hoc community-built data model
Specific and unique for every single project
Linked open data technology (URI-based)
Objects linked via internet protocols (permanent identifiers)
Wikidata simply explained
Triplestore: Hierarchical system (universal database)
Q: Subject: Q-number
P: Predicate, relation: Statement = P-number (property)
Object: Contributing object = Q-number or value
Each object is in turn a subject…
hierarchical = tree structure, recursive
Hierarchy of predicates
Each object is in turn a subject…
hierarchical = tree structure, recursive
Q-number, QID, item number: Unique ID for a Wikidata object
P-number, Property: Wikidata property, relation (extensible)
Statement: Expressing the attribute of a subject using Q/P/V triples
Technology: Wikibase
Wikibase is the software without data…
Developed by Wikimedia Deutschland (2012-)
Forms of Wikibase:
Wikidata datastore
Wikidata is a (non-relational) database
Eventual complete (always work in progress => the world is never finished)
https://www.wikidata.org/wiki/Special:WhatLinksHere/Q106604493
Triplestore
Central place to administer items:
Main property for our purpose
P6104: Maintained by WikiProject
https://www.wikidata.org/wiki/Property:P6104
A Wikidata property allows linking Wikidata items to a WikiProject.
As a consequence, several applications can be served:
How to use it?
Central property: P6104 - Maintained by WikiProject
One item could be handled by multiple Wiki projects and activities.
One item corresponds to one lemma (Wikipedia page)
Possibly on multiple Wikipedia languages (Wikidata sitelinks!)
Images on Wikidata Commons can possibly link to Wikidata item numbers
(SDC = Structured data on Commons, P180 = depicts)
Related properties
Articles are (linked to) items
Create a WikiProject in Wikidata
To be created manually, once:
Example Wiki Project
List of Wiki projects by Chapter
SELECT ?wikiproject ?wikiprojectLabel WHERE {
?wikiproject wdt:P31 wd:Q16695773;
wdt:P664 wd:Q18398868.
SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],mul,nl,fr,en,de". }
}
Gather the list of items for a project (activity, session)
You can often request an extract from the museum catalog system
(often Adlib or successor Axiell Collections)
including additional attributes.
Creative Commons licences
Acronym | Description | Non-commercial use |
CC0 / PD | Public domain | Allowed |
CC BY | Attribution, credits | CC BY-NC |
CC BY-SA | Share alike | CC BY-NC-SA |
CC BY-ND | Non derivative | CC BY-NC-ND |
© | All Rights Reserved | Exceptions for education |
OpenRefine
More information: OpenRefine
Map attributes to Wikidata Properties
For any column you want to register the attribute, you need a unique property.
Maybe there doesn’t exist a convenient property yet?
Then you need to request the creation of a new one.
Example: https://www.wikidata.org/wiki/Property:P10672 (raw material processed)
https://www.wikidata.org/wiki/Wikidata:Property_proposal/uses_natural_resource (approval)
You need to obtain community consensus and approval.
Can be a long and painful process…
not everyone will understand why you need the property…
Upload to Wikidata (example)
Reconcile
Functionality:
Possibility to amend items or create new items
Pay attention not to create duplicates…
Search variant names (synonyms)
First create the subjects = depending objects
OpenRefine reconciliation (example)
OpenRefine Wikidata export (example)
Wikidata Query Project items
wikibase:sitelinks ?linkcount.
wikibase:sitelinks ?linkcount.
SELECT ?item ?linkcount WHERE {
?item wdt:P6104 wd:Q114422535;
wikibase:sitelinks ?linkcount.
MINUS { ?item wdt:P31 wd:Q5. }
}
ListeriaBot
Automatic tool to automatically generate Wikipedia tables,
based on a (very simple) Wikidata Query to obtain a list of item numbers (rows),
adding a (virtual unlimited) list of properties for each item (columns).
Only used in e.g. the Project namespace (never on the main namespace).
Possibility to discriminate between e.g. persons, buildings, or other objects (different properties)
Possibility to segment the items, e.g. by city within a region (same properties)
ListeriaBot usage
Build a Wikipedia target list on Wikipedia in the Project namespace
Example Wikidata Query about people
SELECT ?item ?linkcount WHERE {
?item wdt:P6104 wd:Q114422535;
wikibase:sitelinks ?linkcount;
wdt:P31 wd:Q5.
}
|columns=P18,item,label:name,P735,P734,alias/en:alias,description:description,P937:workplace,P19,P20,P569:birth,P570:death,P27:nationality,P101,P106,P108,P800,P39,P166,P103,?linkcount:pages,P21:gender,P990:voice
Example Wikidata Query excluding people
SELECT ?item ?linkcount WHERE {
?item wdt:P6104 wd:Q114422535;
wikibase:sitelinks ?linkcount.
MINUS { ?item wdt:P31 wd:Q5. }
}
|columns=P18,item,label:name,alias/en:alias,P138,description:description,P101,P10672:processes,P1056,P131:city,P276,P6375,P625:geolocation,P159:mainoffice,P112:founder,P571:created,P576:closed,P1365,P1366,P7888,?linkcount:links
Example sublist by section
Create a list based on a Wikimedia Commons Category
Create a Wikipedia Project page
Describe your project on a project page:
https://nl.wikipedia.org/wiki/Wikipedia:GLAM/Industriemuseum_Gent/Wiki_Café/Fotocollectie_Serck_nv
Use the Campaign dashboard
How dashboard activities are planned (hierarchical 1:M)
Session
event or activity
Physical or online activity
Date - venue
Participants
Subjects
Project page
Documentation
Support
Campaign
List of programs
By institution
By type of activity
By country
Program
List of sessions or activities
By activity type for an institution
Results
Statistics:
Campaign statistics
Activity reporting
Most easy is the Campaign summary report
All items for one Organisation
SELECT DISTINCT ?item WHERE {
?item wdt:P6104 ?wikiproject.
?wikiproject wdt:P31 wd:Q16695773;
wdt:P664 wd:Q18398868.
}
Wikidata Query and visualisation example
Graph of Wiki Projects
Wikimedia Commons photo campaigns
You can include a photo upload campaign:
See:
Wikimedia Commons
Create a photo campaign and/or photo contest
Examples:
Other results
The complete story
You can read more:
Excel > OpenRefine > WikiProject > Wikidata > Wikipedia project page > Campaign dashboard > Wikidata Query > ListeriaBot > Wikipedia > Wikimedia Commons > Wikidata > Campaign statistics
Other tools
Quarry
Advanced reporting
SQL Queries on Dump files…
Pywikibot
Tool to update Wikidata using Python scripts.
Usage:
Prerequisites and techniques:
Two metrics platforms?
There are two platforms to generate statistics
Difference between Outreach Programs and events and the Events Metrics
Similar but different with important overlap?
=> systems could/should be merged? Or is there a specific/intended difference?
Group discussion
Exercises
More reading
Follow-up and coaching
Video chat: https://chat.wikimedia.be
Bluesky: https://bsky.app/profile/wikimedia.be
Questions via: support@wikimedia.be
Credits
Images
https://commons.wikimedia.org/wiki/File:Wikidata-logo.svg
https://commons.wikimedia.org/wiki/File:CC_BY-SA_3.0.png
https://commons.wikimedia.org/wiki/File:Wikipedia-logo-v2.svg
Wikimedia België logo; CC BY-SA; Wikimedia Foundation; https://commons.wikimedia.org/wiki/File:Wikimedia_Belgium.svg
Wikimedia-familie logo; CC BY-SA; Wikimedia Foundation; https://commons.wikimedia.org/wiki/File:Wikimedia_logo_family_complete-2021.svg
MediaWiki logo; CC BY-SA; Wikimedia Foundation; https://commons.wikimedia.org/wiki/File:MediaWiki-2020-logo.svg
Wikimedia Commons logo; CC BY-SA; Wikimedia Foundation; https://commons.wikimedia.org/wiki/File:Commons-logo-en.png
https://commons.wikimedia.org/wiki/File:Wikimedia_Outreach.svg
https://commons.wikimedia.org/wiki/File:5-star_deployment_scheme_for_Open_Data.png
Legal conditions