Could Algorithms Have Saved Warren Beatty?
Forecasting Upsets at the Academy Awards
There is nothing I love predicting more than the Oscars. Nothing. I thought I was pretty good at it, too, until back-to-back upsets (Spotlight in 2016 and Moonlight in 2017) shook my confidence. Until now, I've always relied on intuition, but some people who used data science saw both the 2016 and 2017 upsets coming. Can I repeat their success? And predict the winner for 2018, which is widely considered to be the most unpredictable Oscar race in years? We'll find out ...
Goal: To predict which movie will win Best Picture at the Academy Awards
Dates: October 2017 - present (this project is ongoing and will be updated during the 2018 Oscar season)
Tools: Python (pandas, NumPy, selenium, BeautifulSoup, matplotlib, scikit-learn); classification algorithms (logistic regression, naive Bayes, KNN, random forests, extra trees classifier, gradient boosting, adaptive boosting)
#1 Lesson: Boy, did Lord of the Rings: Fellowship of the Ring get snubbed!! Biggest upset ever. (And, yes, Crash really didn't deserve Best Picture. But you already knew that.)
Where you can learn more: GitHub