Can Big Data predict Oscar Winners? 

Academy AwardOscar Sunday takes place on February 28th. At the end of the night, a few dozen people will walk away as winners, while more than a hundred will smile and say, “It was just an honor to be nominated.”
But while you might be thinking that the Academy Awards have little to do with technology, I beg to differ.

Data miners, like FiveThirtyEight and FarSite, have developed models for determining how the 7,000 Academy members will vote to decide the turnout of the entertainment industry’s most coveted award.

FarSite, the analytics division of IT services provider ICC, has a strong record of success.
They correctly predicted all six major categories in both the 2014 and 2015 Oscars, and 5 out of 6 in 2013. Their model relies on the predictive power of:

  • Wins and nominations from other industry groups, including the Producers Guild of America (PGA), Screen Actors Guild (SAG), Golden Globes (Hollywood Foreign Press Association), Writers Guild, Directors Guild, Critics’ Choice, and British Academy of Film and Television Arts (BAFTA).
  • Prior nominations and award wins.

FarSite also acknowledges that there is some overlap in Academy voters with other guilds—like SAG and Producers Guild, award ceremonies that are presented well in advance of the Oscars. The winner of the PGA (Producers, not Golfers) is considered a strong predictor for the Academy Awards’ “Best Picture” and “Best Producer” outcome at the Oscars.
The PGA awards were presented on January 23, with “The Big Short” winning “Outstanding Producer of Theatrical Motion Picture” for Brad Pitt and his co-producers. “Inside Out” won for “Outstanding Producer of Animated Theatrical Motion Pictures.”
Prior to the PGA Awards, “Spotlight” outshined its competitors as the “Best Picture” favorite, having won the Critics’ Choice for “Best Picture.” “The Revenant” won “Best Drama” at the Golden Globes, and this winner has gone on to take the “Best Picture” award about 50% of the time. In spite of “The Big Short” winning at the PGA, both data science groups are favoring “The Revenant” for “Best Picture” at the 2016 Oscars.
Check out FarSite’s Oscar forecast page here.
FiveThirtyEight follows a similar predictive model.

“We pull from press awards and insider awards. We determine the 25-year rolling success rate of each of the prizes in predicting the eventual Academy Award winner, and square that success rate to determine a score for each award. Films or people who are nominated for a given award get one-fifth of the resulting points, and winner get the full weight. It’s a simple metric, but it’s worked in the past to get a look into the state of the race.”

The entertainment industry isn’t alone when it comes to the use of “big data.”

These large data sets that can be computed and analyzed fuel U.S. presidential campaigns with the insights needed to make rapid changes to gather more votes, as well. In fact, one of my upcoming blogs will talk about the prescience of big data, a growing power tool for predicting presidential election outcomes.
But back to the Oscars…Why not use the voter demographics like they do in the election predictions? Well, unlike America’s registered voters, the Academy’s voters are a mystery.
According to FiveThirtyEight,

“We can’t poll the 7,000-odd people who vote on the Academy Awards. We don’t even have a good idea of who exactly they are. This means that, unlike the methods we use to predict, say, the Iowa caucuses, the problem we’re trying to solve is pretty much stripped of input data.”

Both FarSite and FiveThirtyEight picked the same winners. They also agree there isn’t even a close race in any category.

2016 Oscar Winner Predictions

Best Picture:
“The Revenant”
Best Director: Alejandro Iñárritu (“The Revenant”)
Best Actress: Brie Larson (“Room”)
Best Actor: Leonardo DiCaprio (“The Revenant”)
Best Supporting Actress: Alicia Vikander (“The Danish Girl”)
Best Supporting Actor: Sylvester Stallone (“Creed”)
If you agree with the accuracy of big data, this information might make the Oscars exciting to watch. Of course, there are always a few memorable moments: funny or weepy acceptance speeches, wardrobe malfunctions (and the “what was she thinking?” fashion choices), and host faux pas (Chris Rock emcees again this year).
What do you think of FarSite and FiveThirtyEight’s predictions? And what’s your opinion of big data and data mining for predicting outcomes?
Share your thoughts with us on Twitter @LTronCorp!