Summary:

An MIT study claims that Twitter can be used to predict major social events, thanks to heavy activity from specific communities.

Humanity has long been obsessed with the ability to predict the future, often by the use of magic or clairvoyance. But these days, scientists are trying to predict the future by studying the past, utilizing public records and archives from the places like the New York Times to measure future events based on special cues. Now, an article from MIT suggests that events like the 2013 coup d’état in Egypt could have been predicted simply by measuring activity on Twitter.

Twitter prediction

The report, “Predicting Crowd Behavior with Big Public Data,” focuses on the 2013 Egyptian protests as a model for how to use Twitter to predict major events. MIT Ph.D candidate Nathan Kallus analyzed more than 300,000 open content web sources in 7 di fferent languages and from all over the world, specifically filtering content related to events mentioned in the past and in the future.

From the data, Kallus concluded:

With much of public discourse having at least some presence online and usually more, the wide range of public data captured by our efforts offers unparalleled insight into the futures of countries, cities, and organizations as affected by mass demonstrations and cyber campaigns.

Despite the evidence, there are plenty of caveats to consider. As Kallus mentioned, the Egyptian sentiments leading up to the protests were very strong, and led specifically to sociopolitical event that could be measured. That means that Twitter is not necessarily an accurate predictor of unforeseen disaster — just events of this type. The second is that it’s easy to “predict” the future by extrapolating from the past: because we know that Egypt happened, it’s easy to find evidence that it would happen. Hindsight, it seems, is 20/20.

But what Kallus’s data does suggest is that Twitter is a great way to measure particular social activities in areas of unrest. If the word “protest” gets bandied about heavily among a specific group of people, it’s likely to happen. However, Twitter’s greater capacity for predicting the future is still far from being scientifically proven.

Comments have been disabled for this post