UK General Election 2015

UK General Election 2015

We collected Twitter, Wikipedia, YouTube, and Facebook data on major parties, general hashtags (e.g., #ge2015), all MP candidates whose social media ids are known, and voter advice applications.

The data we produced during our election night data-hack produced a mixed picture. We were able to show that the Liberal Democrats were much weaker than the Tories and Labour on Twitter, whilst the SNP were much stronger; we also showed more Wikipedia interest for the Tories than Labour, both things which chime with the overall results. But a simple summing of mention counts per constituency produces a highly inaccurate picture: generally understating large parties and overstating small ones. And it’s certainly striking that the greater levels of effort Labour were putting into Twitter didn’t translate into electoral success: a warning for campaigns focusing solely on the “online” element.

About

Why is the Internet interesting, when it comes to elections?

The Internet has become a forum where people both find out about politics and manifest their opinions. For example, the UK 2015 General Election has featured in millions of searches on Google, page views on Wikipedia and comments on Facebook and Twitter (to name but a few). This raises the possibility that costly and difficult opinion polling could be supplemented by other measures of what the public are currently thinking about their democracy.

What data did you collect, and what did you do with it?

We collected We collectedTwitter, Wikipedia, YouTube, and Facebook data on major parties, general hashtags (e.g., #ge2015), all MP candidates whose social media ids are known, and voter advice applications. We did some live data-hacking on the night of the election.

Is it really possible to predict elections with social media?

The answer is basically no (at the moment). However, we are interested in exploring the extent to which data from the social web can be used to predict interesting social and political phenomena, especially elections. We think there’s lots of potential in this type of data, but it’s also quite biased and partial: hence most of the results so far have been pretty disappointing in terms of predictive accuracy. In this blog we want to work out what (if anything) data coming from the social web is indicative of. Along the way, we hope to find out something interesting about the way politics and society works.

In Depth