Baidu Uses Big Data to Identify Box Office Fraud
Chinese tech giant Baidu, operator of China’s most popular search engine, plans to release its own sales indices that they say will provide evidence of box office fraud before official figures are released.
Drawing on mapping query data, the positions of their users, the 25 billion location requests they enter, and the Wi-Fi hotspots they log into, Baidu said they are able to actively track the number of people actually watching movies in cinemas.
When movie producers buy out shows to boost ticket totals, the company can draw on its real-time data to verify the number of people attending film showings.
In a report demonstrating the potential of the research, Baidu showed how media reports that the distributor of last years’s box office record-breaker Monster Hunt had made up screenings to inflate box office figures were true.
Baidu’s ability to draw on this real-time data is so unrivaled that it claims it is able to detect fraud before any other party. The report’s authors even call the ability “now-casting” instead of forecasting because they are able to predict same-day box office takings — 24 hours earlier than official statistics.
Wu Haishan, a senior data scientist at Baidu’s Big Data Lab, told China Film Insider that the new box office data may be made available as early as next month.
"We’re going to test more cases about more movies to see if there’s been any box office fraud,” he said. “There have also been other movies that have been accused of box office fraud so we will continue to test those.”
Just this week, official figures released this week revealed a box office slump that is prompting questions about whether the boom has been as big as it first appeared.
The new figures confirmed what private firms like Bejing-based Ent Group had shown a week earlier — that box office takings fell for the first time in half a decade in the second quarter.
The sudden slowdown — after a period of break-neck growth last year in which the country’s box office expanded 49 percent last year — is being blamed by many on a reduction in ticket subsidies made by film producers.
The practice, whereby moviegoers buy tickets at a discount and producers subsidize the remainder of the full price, has led to a distortion in the numbers and lead to doubts about when exactly China will surpass North America as the world’s largest theatrical market.
Other blockbusters such as Lost in Hong Kong and Ip Man 3 have been among a number of domestic movies suspected of inflating box office figures recently.
The company will look towards providing real-time data on current films but that it will require considerable computational resources, Wu said.
Having accurate data about box office takings will not only be useful for investors, but also for the viewing public, he said.
"They’re also very interested in what the real box office numbers are,” Wu said. “Because some producers are using inflated box office numbers to attract more of an audience, but maybe this is not the true case.”