Peeking into the envelopes: Using data to predict the Oscars

Oscardata.nexidia

How historical data allows us to see into the future.

Of all the things that can win awards, movies seem to some of the most subjective. It would follow that the Oscars are relatively impenetrable from predictions as they are largely determined by the whim of some 6,000 movie insiders (how else can you explain Crash?). It seems, however, that as we move more into the world of Big Data, even the fickle minds of the voters can be crunched and quantified into hard numbers and surprisingly accurate predictions.

Around this time last year, Big Data company Farsite began postings results from some data crunching that purported to successfully pick the Academy Award winners from the top six categories; Best Picture, Director, Actress, Actor, Supporting Actress, and Supporting Actor. When the dust had settled Farsite had a perfect 6 for 6 record, beating out the majority of Hollywood insiders.

So how did they do it?

Basically, it’s about knowing your history and keeping an eye on the present. First, Farsite took 40 years’ worth of Oscars history and looked at who at won and what trends were associated with those winners. So, in the case of Best Actor, those trends include things like previous Oscar nominations and how well he has been doing at the other awards shows like the Screen Actors Guild or the Golden Globes. Once the past predictors are taken into account, the next step is too look at what might otherwise be considered the more abstract data; is the popular opinion of the actor positive or has he been involved in some sort of recent controversy?

All that data is then complied for each and every nominated person or film and then dropped into a graph that assigns percentages to each of the nominees. So, for example, Matthew McConaughey was given an overwhelming 82% chance of winning and while 12 Years a Slave squeaked by with a 55% chance.

In the spirit of the season, we here at Nexidia decided to take a stab at our own predictions, both through the tried and true “Googling ‘What will win at the Oscars?’” and some mad data crunching. So, without further ado, are the Nexidia 2015 Oscar Predictions*.

The We-Might-or-Might-Not-Have-Seen-All-the-Movies Marketing Department Picks

Jon Harmer

Best Picture: Boyhood

Director: Richard Linklater

Actor: Eddie Redmayne

Actress: Julianne Moore

Supporting Actor: JK Simmons

Supporting Actress: Patricia Arquette

 

Dana Jones

Best Picture: Birdman

Director: Alejandro G. Iñárritu

Actor: Bradley Cooper

Actress: Julianne Moore

Supporting Actor: Edward Norton

Supporting Actress: Meryl Streep

Kim Junod

Best Picture: American Sniper

Director: Bennett Miller

Actor: Bradley Cooper

Actress: Felicity Jones

Supporting Actor: Mark Ruffalo

Supporting Actress: Emma Stone

 

Thomas Moore

Best Picture: Birdman

Director: Wes Anderson

Actor: Eddie Redmayne

Actress: Julianne Moore

Supporting Actor: JK Simmons

Supporting Actress: Patricia Arquette

 

The Official Nexidia Data Crunched Predictions

 

Best Picture: Boyhood

Director: Alejandro G. Iñárritu

Actor: Eddie Redmayne

Actress: Julianne Moore

Supporting Actor: JK Simmons

Supporting Actress: Patricia Arquette

Interested in the hard numbers? Check out our data here.

*Guaranteed to be correct or your money back

Categories: Best Practices