A new study by a researcher at the University of Illinois suggests that analysing large quantities of world news reports, over long periods of time, could potentially predict the future outbreak of social, political and economic unrest. And can even pinpoint the location of wanted villains like Osama bin Laden.
Kalev Leetaru subjected a massive 30-year archive of worldwide news reports comprising broadcast, print and Internet, to two specific analyses. In the first he looked at tonality of a news report or a blog post, i.e. whether it spoke of a person, place or event in positive or negative terms, using a process called “sentiment mining”. Secondly, he used “full-text geocoding” to pinpoint every news story to a location on the planet which it seems to pertain to. As his study mentions, this is no easy task to do automatically; for example, there are 39 places on the planet called Cairo.
Leetaru claims in his paper titled Culturomics 2.0: Forecasting large-scale human behaviour using global news media tone in time and space that “pooling together the global tone of all news mentions of a country over time appears to accurately forecast its near term stability, including predicting revolutions in Egypt, Tunisia, and Libya, conflict in Serbia, and the stability of Saudi Arabia”.
Taking things one step further, Leetaru claims that geographic analysis of news on Osama could have pinpointed his location to a 200km-wide swathe of land in northern Pakistan, including Abbotabad.
On the face of it, his analysis is a dispassionate and automated version of work that intelligence agencies and think tanks do. Because it is done with a supercomputer, the analysis is devoid of biases or sentiments.
To that extent it seems credible. But so far it has only been shown to work in retrospect. The real challenge will be to see if it works predictively. Is North Korea next? What will happen in Kashmir? Afghanistan?
And what if there simply isn’t enough diverse coverage of small countries and that media anathema—the positive trend?
The greatest lesson from this research could be that while society at large has a tendency to sense negative trends early, it usually seems powerless to do anything. Except blog about it.
Can future events be predicted scientifically? Tell us at email@example.com