Master of Applied Data Science - Milestone II Project Members: Robert Abader, Zealand Cooley, Kendall Dyke
Start date: November 22nd, 2021 Completed: January 31, 2022
When a new movie comes out, reviews from critics often offer some of the first insights as to how “good” it is. These, however, can be hit-or-miss, and so many movie-goers refer to sites like “Rotten Tomatoes” and “IMDB” to look at audience ratings/reviews to get a better sense of what to expect, and whether a particular movie is worth the price of admission. Another perspective is that of the filmmakers since public reception can impact spending on marketing for a film. Could an earlier read on how a film will perform, impact how and where those marketing dollars are spent? The perspective of smaller cinemas is also worth considering here – can theatres with thinner margins benefit by evading “dud” movies? With this project, we explored, with supervised/unsupervised machine learning methods, whether we could predict a movie’s reception based on several key variables and topics associated with each movie.
Throughout this project, we explored movies released in 2020 & 2021. Our main goal was to predict whether a movie will earn a critics’ pick ranking based on several features like runtime, release date, genre, award nominations and wins. To aid in that, we modeled topics associated with plot summaries and movie reviews. We utilized two data sources to aid in our project. The New York Times Movie Review API specified whether the movie received the critics’ pick designation; it also points to a URL containing the NYT review of the movie. To complement this, we used the Open Movie Database (OMDb) API to get more quantifiable information about the movie itself such as runtime, genres, award nominations and wins, number of IMDB ratings and more.
Learn more about our findings here.