Iqra Tariq's ML study reveals genre-time sweet spots for box office hits

A new arXiv paper by Iqra Tariq applies exploratory data analysis and supervised machine learning to quantify the link between movie genre and release timing. Using a dataset of the top 200 box office hits and top 100 flops from IMDb, Box Office Mojo, The Numbers, and Wikipedia, Tariq found statistically significant patterns. Animated and superhero films see their highest success rates (28% and 29% respectively) in June and July. Thrillers and romance genres perform best in November. Conversely, action and comedy films are more likely to flop when released in March, April, or August.

To validate these correlations, Tariq employed multiple regression-based ML algorithms including LWT, Multilayer Perceptron, Random Tree, and Decision Stamp using cross-validation and percentage-split methods. All models consistently reinforced the genre-time dependency hypothesis, achieving high predictive accuracy. The study offers studios a data-driven framework for strategic release scheduling and content production, reducing financial risk in the capital-intensive film industry. It underscores the growing role of analytics in media, akin to its use in finance and e-commerce.