mySociety, Wikidata, Facebook and me: Developing tools that could deepen freely available data on politicians
As one of two newly appointed ‘Community Liaisons’ at mySociety, my role is largely to focus on making Wikidata the leading source of data on every elected politician in the world… just a small task (!). It’s a new role for me and as I settle in and come to know what this actually entails, it is my intention to share (in accessible language) what is happening as mySociety works on this very complex task. So if any of this doesn’t make sense to anyone reading, I ask that you flag it up!
Why are mySociety working on this?
mySociety believes there is a need for a political data infrastructure — a Democratic Commons — that once in place, will give people and organisations more time to focus on holding power to account. This concept comes from the organisations many experiences of supporting people/organisations to build digital tools for democracy. I think it’s a laudable concept and one I am very pleased to be giving my energy to, particularly as mySociety are advocating for the open and collaborative knowledge base Wikidata as the repository for this information.
Enter Wikidata: This post offers a detailed story about mySociety’s journey up to summer 2017 not long after mySociety started to integrate more with Wikidata. From Spring 2017 Tony Bowden and my (brilliant) predecessor Lucy Chambers did a great job of running the ‘Wikidata every politician’ project, thanks to a Wikimedia Foundation grant. This grant has allowed mySociety to support contributions to the project through running Wikidata missions, supported 9 groups internationally to deliver events that improved the quality of data on their politicians, got Tony and Lucy presenting the project at Wikidata con (follow the link to watch the presentation) and allowed for the development of tools that aim to improve the input and management of political data into Wikidata. FYI the tools that have been developed are:
- Compare Wikidata with CSV Allows you to compare information in Wikidata against an external source of data, to highlight errors and omissions.
- PositionStatements This bot adds position held (P39) statements to items, using the same input format as QuickStatements. In contrast to the current version of QuickStatements, this bot can add a second version of a statement which already exists — necessary as an individual may hold a position multiple times and require multiple statements
Working with Facebook: mySociety’s experience from building tools 2 and 3 came in handy when they started to work with Facebook and realised that with a few tweaks, they could develop a tool that could verify crowdsourced data on Politicians from Facebook users and enter a lot of data into Wikidata.
Evolving verification tools: As mySociety draw to the end of the Wikimedia Foundation Grant, they plan to deliver one more tool that can be applied to both Political data and in time, the wider Wikidata/Wikipedia communities. The current thinking is that the tool created for users to verify Facebook crowdsourced data, could be evolved…
Enter ‘Scraper to Wikidata via validity check’ Tool (snappy!): This week, Developers at mySociety started working on an interface tool that will allow potential tool users to verify whether a Politician holds a suggested position (in Wikidata holding this position is called P39) so that this data can be added to the Politicians Wikidata ID and deepen the information known about this person. What is exciting (to me) about this tool is that the data for the suggestions are taken from existing lists/tables of Politicians on Wikipedia. This I think this means that prior to users verifying the position held by the politicians, the Wikidata ID of said politician is already known, so the tool will only ask users to verify Politicians who don’t yet have a P39 statement, or where their P39 statement requires an external reference, such as a link to the government website.
Where the politicians or the party they belong to doesn’t yet exist on Wikidata, the plan is that the tool will allow users to create these items. Thus growing the body of knowledge available on Wikidata!
Tracking the progress of the tool: mySociety developers started working on the tool on 26th June 2018 and I have been attending the daily stand up. The strategy for getting this tool out there is to create a new version of the tool every day! And, I plan to share this process and post more information about what the tool is for, and what it can and can’t (yet) do and its limitations, as often as I can.