On Computational Politics

How big data is reshaping power, politics and the public sphere


Big data helped Germany win the world cup and it let designers create a more comfortable bra. No matter how vaguely understood, the buzzword is full of promise and will add razzle dazzle to any technology article.Big data will give us health, happiness and better bowel movements. According to the MIT Technology Review and, erm, Bono, it will save politics.

Bono is a well known musician

Bono is a well-known musician

Could it also undermine democracy?

Zeynep Tufekci, a researcher in technology and sociology at the University of North Carolina, has written an excellent paper on just this, called ‘Engineering the public: big data, surveillance and computational politics’. Computational politics refers to “applying computational methods to large datasets… for conducting outreach, persuasion and mobilization in the service of electing, furthering or opposing a candidate, a policy or legislation.”

MIT computer scientist Deb Roy created this big data ‘wordscape’ — the landscape that data leaves in its wake

It’s a fairly dense read but a rewarding one. Roughly, Tufekci’s argument is this. We live more and more of our lives online. Those datapoints — an Amazon purchase, a Facebook like — can be mined by interested parties. New computational methods, our much-cherished algorithms, mean that data can be turned into accurate models not just of groups, but of specific individuals.

Combine that with advances in the behavioural sciences, and it becomes possible to enjoy what Tufekci calls “enhanced, network–based social engineering.” (Her italics.) And that social engineering can tested in real time on social networks in real time.

“…these models allow for enhanced, network–based social engineering.” — Zeynep Tufekci

Governing that slightly terrifying infrastructure are opaque or ‘black-box’ algorithms — algorithms that are “proprietary and undisclosed, and most of which are controlled by a few Internet platforms. In other words, an ordinary user sees an opaque algorithm as a black box.”

In fact, we had a glimpse inside that black box recently, when Facebook published a paper on how it manipulated users’ emotions, and it scared us. There was a huge and justifiable outcry. But Facebook — like all mass media — manipulates us all the time. In this instance, we actually got to see how they did it. It’s easy to see how such tools could be used for political manipulation.

And worryingly, this sort of networked political manipulation suits those already in power — as Tufekci puts it, “incumbents who already have troves of data, and favors entrenched and moneyed candidates within parties, as well as the data–rich among existing parties.”


This concern isn’t merely academic. Computational politics will play a key role in the 2015 general election, as this article sets out. And the principles and mechanisms behind computational politics are being used in many other fields, from business to finance to medicine. It’s pretty much how the consumer internet works, from Amazon to Facebook.

So what do we do about it? Tufekci calls for more research and regulatory oversight. We actually tried something like that here in the UK, with the midata initiative, which started in 2011. It aimed to give consumers access to the data companies hold on them but was shelved.

Midata is due for review at the end of this year. The idea, if not the implementation, remains a good one. Start ups are also wondering if there could be businesses built around returning control of personal data.

In the new age of computational politics, we’ll likely need both approaches.