Trumped by Data: Can We Predict the Choices of the People?

Data and the US election

By Emma Sampayo

With this week’s electoral results culminating in Trump’s victory, many of the polls got it wrong when it came to predicting who would win. As stated in an article in the New York Times, “Donald J. Trump’s victory ran counter to almost every major forecast — undercutting the belief that analyzing reams of data can accurately predict events. Voters demonstrated how much predictive analytics, and election forecasting remains a young science: Some people may have been misled into thinking Hillary Clinton’s win was assured because some of the forecasts lacked context explaining potentially wide margins of error.

There are many types of data that can be used to understand what political view point people have. Many companies are hailing social media as the new best thing to use to predict future votes, but is social media a good source of data for predicting how people are going to vote? After all, there are far fewer people on Twitter than in the US electorate so is it really representative? And there are many other social media sites, so are all opinions the same for one person across all their social media platforms? It could lead to not being a true reflection of how people feel; again lacking context.

Alas! One may cry, data is data and people are expressing what they feel and think on social media sites every day, every hour and every minute. But can we use this data to predict future results? Future elections?

Possibly.

Earlier this year, the pollsters where shocked to see the ‘voice of the public’ stand affront as the UK won the vote to leave the EU; (some ‘remainers’ are still just as shocked). Again, many people did not expect this to happen, leaving the pound to plummet to levels not seen for nearly 30 years. So, what is being used to predict these future trends? Is predictive analytics being carried out and what could be done to better our chance of knowing what can happen?

Research suggest that human behaviour is hard to predict. Yes, you can predict generic profiles, ‘women in their 30’s’, ‘men over 50 are more likely to’ etc – but to actually predict what is going to happy to one person, at one time, is hard as there is always an element of chaos and free will. Buying patterns can be predicted – socials scenes – yes – however individual behaviour is hard because of spontaneity.

Do we need artificial intelligence to lead the way in understanding human behaviour? Maybe. But how?

There is a company called Sandtable, where a couple of S2DS  Fellows work, that believes in Agent Based Modelling – a simulation method to understand human behaviour under different circumstances, either groups or individuals – which could help, but I won’t go into too much detail about AI now.

And what about Trump? Well, that’s a whole different kind of intelligence, isn’t it?