This repository contains code used to generate the paper Box-Office-Case-Study.pdf.
Movie production companies are interested in understanding variables contributing to a successful film, particularly whether film budget and IMBD score are predictive of net profits. Box office data from all 2019 film releases was collected and enhanced with additional features to address this question.
Initially, a linear model was used to explore response associations and establish a baseline. The initial model displayed poor fit according to the residual standard error distribution, warranting a more complicated analysis. Several hierarchical models were used to address heteroscedastic variance concerns and ultimately a mixed effect model was applied to observations based on release month, which enhanced estimation accuracy for most fixed effects.
Final estimates for 'Budget' and 'Critic.Score' have statistically significant positive effects on profitability at the 95% level. On average, every $1 increase in budget yields an additional profit of $2.43 and improving the IMBD score by one point leads to a net profit increase of ~$45 million. In addition, augmented field 'Title.Sentiment', defined as the polarity score of the film's title, showed surprising significance, indicating a ~$97 million net profit increase per unit gain of sentiment score.
For production companies, it's not always feasible to create high budget, critically acclaimed films, but they can control the naming. One actionable recommendation is to title movies positively, as audiences seem more inclined to purchase tickets to positively named films.
Box-Office-Case-Study.pdf
: Written report, outlining data analysis procedure, discussion, and resultsBox-Office-Case-Study.Rmd
: Code used to generateBox-Office-Case-Study.pdf
./data/
: Raw data directory