If you’d asked me back in 2005, I’d have told you that the 2010 election would be the first online election. It turned out not to be.

For example, the YouTube and Facebook leaders debate was much less important than the Television debates.

However, there are a few places relating to data where the Internet did something genuinely new this time.

MP candidate data

The most basic data about an election is the names who you can vote for. Shockingly, there is no central, official source. Never mind one with an open data license. The data from newspaper websites is usually incomplete, particularly for independent candidates.

In the UK, we have Parliaments of an irregular length. This means that officially, you only know the list of candidates 2 weeks before the election. Before that you have to make do with screen scraping party website lists of Prospective Parliamentary Candidates (the obtuse term for someone who is going to be a candidate, before they can officially be one. e.g. the Conservatives) and hoping to get data from Wikipedia.

This year Edmund von der Burg bravely overcame all these problems with his YourNextMP, which he has run entirely as a volunteer. Not only does it have the names and parties of candidates, but it has extra information like email addresses, photos, schools attended and web site links. The data is available under a CC BY-SA license.

After nominations close, just 2 weeks before the election, each local authority publishes notices of poll (e.g. Liverpool, Riverside). They are the official list. Amazingly, the volunteers building up the YourNextMP data set cross checked their data against 650 notices of poll, in just a few days after nominations closed.

Really, all the basic work of YourNextMP should be done by the state. We could have fixed term Parliaments, with nominations that close two months before the election. The local authorities could upload basic candidate data, including electronic and paper contact addresses, to a central website, perhaps run by the Electoral Commission.

Then YourNextMP could concentrate on the added value. What we ought to be doing – researching the candidates to find out more about them. What companies have they been directors of? What charities do they support? Will they voluntarily declare their interests in advance of the election (as recommended by the Ministry of Justice)?

Maybe candidates should have to declare a bunch of “same as” RDF identifiers – such as their unique codes in the companies house database, the land registry and Wikipedia.

Julian Todd from the Straight Choice thinks every candidate should be obliged to publish a full CV, perhaps as structured data (see last paragraph of this article). And why not? Currently, we ask for far less of our new employee in Parliament than we would of somebody we employ in our business.

Election leaflets

As I said, this election was another offline election. Part of that is the mass media, big leaders. But the other key part is getting out the vote. It is the door to door canvassing, the hard labours of local party workers up and down the country. Vital are election leaflets, a data set hitherto hidden from us.

The Straight Choice has crowd sourced 5173 election leaflets, from all parties and most constituencies (disclosure, I do some systems administration for them). You can see a zoomable map of them, and a mosaic of the party leaders made of their leaflets, in this blog post where they report back on what they’ve found.

Have a read through the presentation at the end of that blog post. The Straight Choice have a series of campaigning demands. They’re all data related.

As I said above, they’d like CVs of candidates. If we continue to have a non-proportional electoral system, they’d like local voting intention polling – essential data to properly tactically vote. And finally, they want every electoral leaflet to be sent to the Electoral Commission and published. Like a copyright library, so electoral law can be properly enforced.

Just like YourNextMP, The Straight Choice is run entirely by volunteers. Julian Todd and Richard Pope did the central work.

Please please please, upload any leaflets you have – it’s vital to catch lots in the “end game”, as they can be particularly dirty.

Candidate opinions

Wouldn’t it be nice to have structured data on what the candidates think on a series of local and national issues? Luckily some volunteers, along with a small charity, found out using an incredibly complicated crowd sourcing operation.

The hinge of this was Democracy Club, a network of over 6000 transparency activists in nearly every constituency in the country. It’s amazing what you need to build, when you don’t have handy JSON files.

Once again, Democracy Club was started and is run by volunteers – Seb Bacon and Tim Green. For the last few weeks, mySociety got an emergency grant from the Joseph Rowntree Reform Trust to pay Seb, so Democracy Club could do even more in the run up to the election. Thank you to them!

The Democracy Club members built up the YourNextMP database of candidates, and uploaded lots of the leaflets to the Straight Choice. They also made a database of local issues in each constituency. These were munged together by mySociety (who I work for), into a survey to all candidates.

You can view the results for your constituency by entering your postcode on TheyWorkForYou’s election site. Please pass it around.

(By the way, the data for even that postcode lookup caused complications because the election is fought on new constituency boundaries. Matthew Somerville from mySociety worked them out and offers an API, although some political parties have had trouble.)

You can download the candidate survey response data from the
TheyWorkForYou Election API.

Election results

Finally, by the time the counts are finished on Friday morning, you’ll want to find out who won. There are two elections happening tomorrow. One for Westminster, but also one for local councillors in your area.

Chris Taggart, another volunteer, has come to the rescue with his Open Election Project. He’s been promoting an electoral results RDFa – a neat, lightweight form of the semantic web that embeds in extra HTML tags, so it is easy for councils to add them in their existing content management system.

He’s persuaded quite a few councils to start publishing their data in this format, and invented a new technique of asking ‘are you an enabler or a blocker?‘. Hopefully in a few years time, he’ll have got every council to publish data in this format.

The learning process will have taught everyone how to make progress on the difficult question of going from the theory of national, open sets of local data, to the practice.


Have a great General Election!

+ posts

CEO of ScraperWiki. Made several of the world's first civic websites, such as TheyWorkForYou and WhatDoTheyKnow.

6 thoughts on “Election data!”

  1. One awesome thing about all these is that the data is reusable. It’d be great to know how many second-tier projects have been enabled by things like YourNextMP and the Straight Choice (and not just because I made one).

  2. “Then YourNextMP could concentrate on the added value. What we ought to be doing – researching the candidates to find out more about them.”

    It would be great to link those data in real time to candidates appearances on TV, the way I imagined at http://stop.zona-m.net/node/64

    that is, correlate TV-listings downloadable from the Internet to these websites in such a way that if you watch a TV show on the PC, you can have automatically a browser window opened on all the history of the candidates presents in that show, so you can check in real time if they are coherent with what they said earlier etc…

Comments are closed.