Wikidata:WikidataCon 2017/Notes/Wikidata: State of the Project

Title: Wikidata: State of the Project

Note-taker(s): VIGNERON

This is the Etherpad for note-taking of this session.

Speaker(s)

Name or username: Lydia Pintscher

Contact (email, Twitter, etc.): https://twitter.com/nightrose

Useful links: slides: https://commons.wikimedia.org/wiki/File:WikidataCon_2017-_State_of_the_Project.pdf, https://www.wikidata.org/wiki/Q18016466 ...

Abstract

Wikidata is turning 5 years old. A lot has happened since it started as an empty wiki with no items, properties and statements. We'll take a look at the current state of the project and see where the next years might take us. What data do we have? In what state is it? What does our community look like? Who is using our data to build amazing things? Where do we still have work to do? What are the biggest opportunities and challenges ahead of us? I'll try to give some answers to these and more.

Collaborative notes of the session

5 years, yeah ! Who is here since the beggining (raise hand), who is here for less than a year (please help them to discover)

1400 very active editors

37,8 millions items (especially created in the last months)

3900 properties

various stats from http://tools.wmflabs.org/wikidata-todo/stats.php

more and more items not linked to any wikimedia projects, good things to being a project of our own ;)

A lot of infoboxes

+60% usage

8.5 millions SPARQL query per day!! (big thanks to the team behind it).

1/3 of all edits on Wikimedia projects are on Wikidata

+2% growth of non bot edits on Wikimedia projets, and almost all of that is on Wikidata

UX improvements - first citoid support -

...

Examples of good queries (plants with emoji, moon seed trees location, etc.)

ArticlePlaceholder, it helps support more languages!

Structured data Commons, yeah! Commons is a treasure but a buried one.

Freebase is gone :/

Games like Guessr

Scholia : https://tools.wmflabs.org/scholia/ (presentation later today)

Monumental, maps about monuments: https://tools.wmflabs.org/monumental/

Gender gap on Wikidata http://tools.dicare.org/gaps/gender.php

2016 wasn't that terrible, buzzfeed use Wikidata data for an article!

Making Wikidata more useful inside and outisde Wikimedia (inside : Wiktionaries, etc. outside : partners, allies, make it easier to install, federation, etc.)

Questions / Answers

• Q: Geertivp: Did you think about making constraints checking proactive/mandatory?
• A: Lydia : I think we should not be too restrictive, there is too much complexity and exception in the world. For example, there's a constraint for people that cannot be born AFTER they died, but think of a fictional timetraveler... :) (thankfully there is none *yet* in Wikidata :P )

• Q: Geertivp: Do you think to help identify wrong/not true statements?
• A: Lydia: there is a qualifier "disputed by" that you can use for this.

• Q: [sabas88, remote] Is there a plan for a mobile / responsive interface to be developed?
• A: Not at the moment. We need to rework most of the input part of Wikidata before working on UI, so this is what is holding us down for the moment. I don't plan this will happen in the near future.

• Q: [sabas88, remote] Instead of creating a property for each catalog (Pxxxx), can we assign to the catalogs a new element (like Cxxxx)?
• A: This is actually the fist time I hear about this, please ask the user to get in touch. [I explained it better to good guy Sannita]

Luca Martinelli read questions from the etherpad! Luca is a good guy, be like Luca! You're welcome <3

• Q: Asaf : who are the culprits for mass import ?
• A: [Daniel Mietchen] Mostly scientific papers, we're going to have a panel about this later, WikiCite session.

• Q: What kind of transparency is there for automatic checks and constraint check.
• All rules are public, defined in the property itself ; anyone can see them and edit those if needed

• Q: Wikibase, when it come to encourage people to install it ; what about others useful (almost mandatory) tools that goes with it?
• A: Yes, I think so. This is something that I want to address in the next year, for now it's still too hard to install Wikibase and it shouldn't be so hard.

• Q: Performance isse like TimeOut on Query service (but not only there)
• A: depends on the tools, for specific need go to specific person (for Wikidata Query answer by ...), some queries ill always be too big and expensive

https://commons.wikimedia.org/wiki/Category:Wikidata_visualizations

• Q: Andy: copying labels could be a solution, but there is some resitance, as it could/should be deal be the software and not by bots
• A: Lydia: there is the fallback, it help a bit. Copying is not necessary. And I don't want to encourage to game this just to improve the statistics ;)

• Q: Magnus Sälgö (User:salgo60) Can we work on getting better trust about facts using "weekly checks" that the Wikidata facts are consistent with "trusted" sources and that we document differencies? Plus quality of external sources should be more transparent to the readers.
• A: ... we should work on this but I don't know when

• Liam: ....
• ....

• Harmonia: answering about labels copying, there is project for name and it can't be done fast as it should be done in a clean way, please don't gamify it!