Twitter offers fantastic insight into what’s happening right now, but can it also be used as a crystal ball to see into the future? That’s the question a recent study may have some answers to.
As MIT Technology Review reports, researcher Nathan Kallus of MIT believes he has come up with a method of predicting crowd behavior offline based on tweets.
His research paper Predicting Crowd Behavior with Big Public Data explores the Egyptian coup d’état in 2013 and how protesters used Twitter to coordinate their meetings.
Kallus’ model mines Twitter data for mentions of specific future events, and analyses trends around these mentions. He used data from Recorded Future, a web intelligence software company, to make his predictions.
The paper shows that there was significant indications that Egyptians would protest in advance of the anniversary of President Morsi’s rule, and that the data went further to show that the protests would last longer than the anniversary itself.
However, one of the biggest questions about the usefulness of Kallus’ new method is whether it could be used to predict an event before it happens, or if it relies on the power of hindsight to predict events in history.