Social media predictive power might also extend beyond box office success

Twitter Box Office Prediction Will titans clash for "Clash of the Titans?" Let's ask Twitter

Tapping into the wisdom of the crowds to forecast future trends has served prediction markets well for years, but Twitter might be even more effective than even the biggest and most widely used market, the Hollywood Stock Exchange. In a recent study, watching tweets among Twitter users allowed HP Labs researchers to predict box office figures better than the Hollywood Stock Exchange, Fast Company reports.

Sitaram Asur and Bernardo Huberman of HP Labs kept track of movie mentions on among 2.9 million tweets from 1.2 million users for three months. Their sample focus included 24 movies such as box office king Avatar and Twilight: New Moon.

For opening weekend, their computer model monitored the rate of tweets near a movie's release date and also factored in the number of theaters showing the flick. That allowed the model to predict the opening weekend revenues with 97.3 percent accuracy, compared to the Hollywood Stock Exchange's 96.5 percent accuracy.

For the second weekend, the model examined both tweet rates and the ratio of positive to negative tweets. That different approach reflects second-weekend performance success based on word-of-mouth, rather than opening-weekend performance buzz. Again, the model delivered quite splendidly with 94 percent accuracy.

There's a few caveats to keep in mind, such as the fact that the Twitterati represents a certain self-selecting slice of society. One expert told Fast Company that he suspected Twitter might do better for predicting "upmarket" films aimed at somewhat older audiences, given the social media service's user base.


But there's an exciting possibility for extending Twitter's predictive power beyond movies, the researchers say. A similar model could apply for any number of commercial products beyond Hollywood fare, and might work especially well for products or trends that lack prediction markets such as the Hollywood Stock exchange. Presidential election results, anyone?

Twitter might even go beyond a mere forecasting service for commercial success. Technology Review suggested that the social media service could also allow marketers to directly influence the success of their product by boosting tweet rates -- although we assume that strategy would require more savvy than using obvious Twitter bots.

There's also a good opportunity here for savvy prediction market players, if they can apply a bit of Twitter analysis to how they play the Hollywood Stock Exchange. We hear that HSX has recently transitioned to real money.

[via Fast Company]

3 Comments

Hey Jeremy - I work with Bernardo and his lab at HP. Want a chat with him to learn more about this and other research? Hit me up at matt dot mclernon at bitepr dot com.

I think that Twitter and other social networking sites could become super-large tools for social scientists and other researchers in the coming years. Two days ago, I read a story from NewScientist about a man named Pete Warden;

Google "Data sifted from Facebook wiped after legal threats"

He took data from millions of Facebook profiles and constructed a map of the US with different regional clusters, how we make friends in each of them, and showing what people in each region are like. Though the data is now gone (read the article), I think that he found how Internet tools like this could reveal tremendous discoveries about our society.

I only wonder what may come to be known using these tools. This is just the beginning.

Jeremy,

I've recently been working on a similar concept - I call it "Local Consciousness"

With the advent of Twitter and location based services (LBS's) we are projecting our thoughts and actions into the public domain. We share our plans, location and actions in 140 characters or with a simple check-in. This is essentially data.

By giving the value of each tweet or check-in a binary value of either 1 or 0, we can plug those results into a simple correlation statistic and measure the results against stimuli or other verticals. This information can be sold to restaurants, on the short term, to change offers or push certain products. On the long term it has relevancy for city planning, business planning and sales forecasting.



June 2013: American Energy Independence

Five amazing, clean technologies that will set us free, in this month's energy-focused issue. Also: how to build a better bomb detector, the robotic toys that are raising your children, a human catapult, the world's smallest arcade, and much more.


Online Content Director: Suzanne LaBarre | Email
Senior Editor: Paul Adams | Email
Associate Editor: Dan Nosowitz | Email
Assistant Editor: Colin Lecher | Email
Assistant Editor: Rose Pastore | Email

Contributing Writers:
Rebecca Boyle | Email
Kelsey D. Atherton | Email
Francie Diep | Email
Shaunacy Ferro | Email

circ-top-header.gif
circ-cover.gif