And the winner is … a Yale graduate student with a system for picking the Oscars

A Yale chemist is earning rave reviewers for her ability as an Academy Awards prognosticator.

Allison Walker, a Ph.D. student in the lab of Alanna Schepartz, was featured several weeks ago on the website FiveThirtyEight, in a story about people who have developed their own statistical models for predicting who will win Oscars. As it turned out, Walker’s approached proved quite successful.

Walker correctly predicted winners in five of the six major categories. She was on the mark with the awards for best director, actor, actress, supporting actress, and supporting actor. Her only misstep was best picture; she chose “The Revenant,” rather than “Spotlight.”

“I was kind of hoping I would get them all right, but I still did better than I expected,” Walker said the morning after the awards broadcast. “I’m pretty happy with it.”

In the FiveThirtyEight story, Walker said she was developing a system that examined movie reviews for nominated films. It would look for words that are statistically associated with winning an Academy Award, and come up with a probability for winning the award. In the case of acting categories, it would look for winning words appearing in the same sentence with an actor’s name.

Walker said she’s been contacted by FiveThirtyEight to do a follow-up story, and at least one local TV news station wants to interview her.

She also noted that her system could have applications beyond the Oscars. For example, she said, the system could be used to predict music or television awards. It might even be applied to social media posts about politics, to predict voting trends.

 

Oscar image via Shutterstock