Friday, February 10, 2012

#ALGORITHMS: "People's Oscars Gauge Public Sentiment"

The People's Oscars project intends to demonstrate that business analytics and natural language software can be successfully used to understand, respond, and predict public sentiment in Academy Awards winners, sports events, retailing trends, and journalist media.

The Academy Awards each year bestows Oscars to the judges’ choices for Best Actor, Best Actress, and Best Picture. These choices often do not jive with the actual popularity of the characters and movies themselves.

The People's Oscar seeks to remedy that by gauging public sentiment itself. However, instead of asking people to log-on to a website and explicitly vote for their favorites (thereby limiting the sample size to activists who take the time to vote), the project uses IBM analytics and natural language understanding algorithms to extract the people's real sentiments from their tweets.

The People's Oscars, sponsored by the Annenberg Innovation Lab at the University of Southern California and the Los Angeles Times newspaper, in addition to IBM, aims to retarget Watson-like analytics to measure sentiment from social media postings. Just as Twitter, Facebook, and FourSquare have changed the way consumers communicate, USC, IBM and the LA Times are seeking to change the way Oscars are bestowed--at least in their cyber-space incarnations.

Now that a billion of users worldwide are using social media to communicate, their sentiments regarding the best actors, actresses, and movies should theoretically represent public opinion. However, to mine that vast deposit of online data requires high-powered analytics that was previously only available to enterprises. The partnership between USC's Annenberg Innovation Lab, IBM, and the LA Times intends to demonstrate that business analytics can be successfully used to understand, respond, and predict public sentiment in applications from sports forecasting to retailing to journalist media.

This collaboration "demonstrates how the media industry is advancing at a pace consistent with the fast-evolving Twitterverse," said Steve Canepa, General Manager of Media and Entertainment, IBM. "By gaining insight from the growing world of social media, we can add a new layer of intelligence that will change how all industries engage with their digitally savvy consumers."

Key to the success of The People's Oscars was repurposing IBM's business analytics and natural language understanding software to extract pro and con sentiments from the 140 character tweets of the Twitterverse. By crafting algorithms that can detect the difference between genuine admiration and mere sarcasm, IBM's natural language understanding software was able to gauge the popularity of actors, actresses, and the movies themselves from millions of tweets.

Analytics "can capture valuable information and opinions derived from movie fans," said Professor Jonathan Taplin, Director of the USC Annenberg Innovation Lab. "The People's Oscars moves beyond the pundits' opinions of who the winners may be, to understanding who real moviegoers want to see receive the highest accolades."

The latest picks and pans in The People's Oscars are being published daily in an infographic illustrating ongoing sentiment in the LA Times leading up to, the official 84th Academy Awards ceremony where the Oscars will be presented on February 26, 2012.