Can We Datafy Human Behaviour?

Can we predict what humans will do or is still science fiction?


Recent years have heralded a growing influence of Big Data and analytics. This has cut to the core of what it means to be human in the very realist sense, with the ability of computers to track what we do, who we talk to and even what our future actions may be.

The big questions that surround this data are how it is collected and how it is analyzed.

Collection of this kind of information has never been easier than it has been today. In the past in order to get sentiment it was done with focus groups or based on the use of surveys to get public opinion on a subject or product. Today this kind of sentiment can be found almost instantaneously through the use of sentiment analysis on social media.

For individual objects this is essentially a case of crowd sourcing. It is possible to create keywords associated with the product, which often give a strong indication of the general perception. For instance, social media was one of the key indicators of the success of ads at the Superbowl where Twitter was mined in order to get the general sentiment of those who watched it.

With this kind of analytics it is not only useful for looking at what the group thinks, but through analyzing the patterns of like/dislike for an individual a fairly detailed picture can be built on them that allows companies to predict what they will like and dislike.

Equally we can see through the datafication of books, images, films and music, plus the sentiment expressed about them, that patterns can be created that can then be exploited. The best example of this is Netflix’s use of analytics and data to create House of Cards. Here they saw through data that certain people liked political dramas and Kevin Spacey. Through putting these two together, combined with other metrics, they created a hit show.

Essentially Netflix showed that data could predict that many people would want to spend hours watching this show.

As more of this data is collected, connections made and correlations between data sets identified, the opportunity to predict human behaviour becomes even closer to reality.

But is this really data and machines predicting human behaviour?

The reality is far from this, as the computers and data simply show the answers to questions asked by humans. For computers showing words expressing a dislike to something is the same as showing numbers of crop yields. There is no distinguishing between the two by a machine as it is simply displaying the data that it has to the person who has asked for it.

House of Cards came about not because a computer showed somebody that people like Kevin Spacey, it was because a human put the two things together and made the show. It was because a human wanted to see which actors people liked and which genre intersected this.

So can we datafy human behaviour? To an extent we can, but the reality is that there is a distinct line to be drawn between collecting the data on how we act, then the way it can be used. There is no way that internal emotions can be tracked or how somebody wants to portray themselves online. It could even come down to a muddling of two human beings using one shared computer.

So although we can predict certain things, the chances of computers being able to predict what we do, with who and when are still as far away now as they were 20 years ago. Luckily for us, Minority report still seems like a work of fiction, regardless of what many within other media outlets would have you believe. 


Read next:

Working At The Boundaries Of Aesthetics And Inference