Reading all 1.2M films in the TMDB catalogue with R: data cleaning at scale, custom ggplot2 visual design, a predictive model, and a hard look at the biases (survivorship, review-bombing, missing data) that fool most movie analyses.