The 94th Academy Awards honoring movies released in 2021 will take place on March 27, 2022. This year, we’d like to make our annual ritual in predicting the Oscar winners via Machine Learning even more engaging by inviting all of you to actively take part in this fun challenge.
This dataset is prepared to predict the following 20 categories. Feel free to predict all of them, any of them, or just one. It’s up to you! Just know that you will probably need to build one model per category for better results.
- Best Picture
- Best Director
- Best Actor in a Leading Role
- Best Actress in a Leading Role
- Best Actor in a Supporting Role
- Best Actress in a Supporting Role
- Best Adapted Screenplay
- Best Original Screenplay
- Best Cinematography
- Best Costume Design
- Best Film Editing
- Best Sound Editing
- Best Sound Mixing
- Best Visual Effects
- Best Makeup and Hairstyling
- Best Music, Original Song
- Best Music, Original Score
- Best Production Design
- Best International Feature Film
- Best Animated Feature Film
The following three datasets include relevant information about movie industry awards celebrated up until March 20, 2022. Feel free to use these datasets or add other data sources as you see fit.
- Combined dataset (also shared above) with movies 2000-2021
- Training dataset with movies 2000-2020
- Test dataset with movies 2021
Kindly notice that the objective fields are listed as “Oscar_Best_XXXXXXX_won” and there are 20 different categories you can predict. You will most likely need to build one model per category. The remaining fields are the predictor fields ,e.g., “Oscar_Best_XXXXXX_nominated”. Below are some explanations.
- XXXXXXX_won: Numeric (number of awards won at that specific event)
- XXXXXXX_won_categories: Items (categories that were won at that specific event)
- XXXXXXX_nominated: Numeric (number of nominations for that specific event)
- XXXXXXX_nominated_categories: Items (nominations for that specific event)
The deadline to submit your predictions is March 23 at the end of your business day. You can submit your results by sending an email to email@example.com.
You can submit a maximum of two picks per category in case you are using different approaches, e.g., ML model-based prediction vs. your intuition. Please stick to the following format when submitting.
BigML Username: ExampleUser
- ML-based: Dune (OPTIONAL: X% prediction confidence. Its not mandatory to use BigML, however, if you do you can share your BigML model via a secret link too.)
- Intuition: Don’t Look Up
We want the world to know how good your predictions are so we’ll present the best performing predictors on BigML’s social media channels as well as our blog. Good luck!