HP Labs Use Twitter to Predict Box Office Success: 97.3% Accurate

by Jacob Zinn | April 9, 2010 at 03:31 pm

368 views | 0 Recommendations | 0 comments

Researchers are using Twitter to predict box office performance with 97.3% accuracy based on how many tweets are posted hourly about upcoming films.

Between November 2009 and February 2010, researchers Sitaram Asur and Bernardo Huberman at Silicon Valley’s HP Labs tracked 2.89 million tweets from 1.2 million users about 24 movies released in the 3-month period to accurately project opening night earnings.

Researches Use “Tweet Rate” for Box Office Success

They determined that the hourly number of tweets (or “tweet rate”) could give estimates of how well or how poorly a film would do when it hit theatres.

There is a strong correlation between the amount of attention a given topic has [in this case a forthcoming movie] and its ranking in the future,” wrote Asur and Huberman in their paper, Predicting the Future With Social Media.

Based on tweets, HP Labs predicted The Crazies would earn $16.8 million in its opening weekend in the U.S. It took $16.06 million.

They also predicted Dear John would earn $30.71 million in its opening weekend in the U.S. It took $30.46 million.

Tweet Sentiments Predict Box Office Performance

HP Labs also looked at the number of tweets that spoke positively or negatively about a movie and based part of the performance on that.

We developed algorithms to analyse these tweets and measure the rate at which they were produced,” said Huberman. “Our intuition was that the faster people tweet, the more likely they are to go and see it.

After The Blind Side’s opening weekend that earned the film $34 million, viewers tweeted positively about the film and it made $40 million in its second week.

On the other end, New Moon’s box office dropped by more than 66% from its opening weekend, which correlated with negative tweets following the vampire flick’s release.

The researchers feel the same algorithm could be used to predict outcomes of political elections and stock market drops, among other things.