According to computer scientists, the Microblogging site Twitter can be used to predict which films are going to be blockbusters and which aren’t and it’s apparently turning out to be more accurate than other predictive tools such as the Hollywood Stock Exchange.
The scientists involved studied around 3 million Tweets with content that covered about 25 different films, one of which was Avatar of course, and found that the rate of messages produced was a good indication of the success of the movie.
“Our predictions were incredibly close,” Bernardo Huberman, head of the social computing lab at HP, told BBC News.
He said the system predicted that the film ‘the Crazies’ would take $16.8m in the first weekend it opened in the US whereas it took an incredibly close $16.6m.
Same story for the film ‘Dear John’, the system predicted it would take $30.71m and it actually took 30.46m.
You’ve got to admit that’s about as close as you could expect to get with the most sophisticated of methods, and without knowing tickets sales in advance.
The researchers got their figures by analysing the rate of tweets in the run up to the release of a move and then analysed the content of the tweets after release.
“We developed algorithms to analyse these tweets and measure the rate at which they were produced, said Dr Huberman.
“Our intuition was that the faster people tweet, the more likely they are to go and see it.”
So that’s the initial success, in order to asses the likelihood of ongoing success, the researchers conducted what they called “sentiment analysis”.
The research team used Amazon’s Mechanical Turk for this, this is a method whereby real people are paid to perform some tasks that computers would struggle to complete and in this case they decided if the content of the tweets were positive, negative or neutral.
“It’s tapping into collective intelligence,” said Dr Huberman, who carried out the work with Sitaram Asur, of HP.
However, according to Huberman, Twitter users tend to be young, tech savvy and reasonably well off so may not be a suitable demographic for some trends, but it could be of use in forecasting how well a gadget or product will sell.
If you want to read the full research it’s available on the Arxiv website.
