Movie executive bet millions of dollars on blockbusters, cross their fingers and hope the movies will do well at the box office. They’re always looking for signs, like the pre-release buzz around a movie, that their bets were right. But how do you quantify buzz? One group of researchers suggests looking at Wikipedia edits.
The premise of the study, published in PloS ONE, is that the more people who are editing and updating a movie’s Wikipedia page, the more people are interested in that movie, and the more people will go see. it. The researchers tested this hypothesis by tracking the wikipedia pages for 312 movies that came out in 2010. They estimated the popularity of each page by combining views, number of users, number of edits and how rigorous the chain of edits were. They then compared that to the predicted box-office revenues. It turns out that their Wikipedia algorithm could guess the success of a movie with 77 percent accuracy. And the more successful the film was, the more accurate the algorithm was.
The author say that this method of prediction doesn’t have to stay in the movie realm. “The introduced approach can be easily generalized to other fields where mining of public opinion provides valuable insights, e.g., financial decisions, policy making, and governance.” Perhaps soon instead of buying Twitter bots, politicians will be buying Wikipedia editors.
More from Smithsonian.com: