Letterboxd at the Movies
A personal statistical analysis of films released 2020+ comparing Letterboxd watch counts to box office performance, identifying what features predict appeal to the Letterboxd audience.
Universe
Find a film
Top 20 by per-segment residual
The project's headline metric. Films punching most above their weight on Letterboxd, controlling for theatrical reach and release year within their own release class. Positive residual = more LB-watched than the within-segment regression predicts. See methodology for caveats.
Distribution of residuals
Mean-zero by construction (the regression includes an intercept). Most films cluster within ±1; the tails are where the leaderboards live. Red line = expected (residual 0).
Top 3 per segment
Preview of each segment's leaderboard. Click a segment header for the full ranking.
Download
About the data
- Universe: union of TMDB discover (vote_count ≥ 25) and Box Office Mojo yearly charts, deduped by TMDB ID.
- Letterboxd metrics: watches, likes, average rating, rating count, themes — scraped per film via Playwright (Cloudflare-mediated).
- Box office: Mojo for worldwide/domestic gross; The Numbers fills budget gaps where TMDB lacks them.
- Residuals: per-segment OLS fits of
log(watches) ~ log(countries_released+1) + log(months_age) + C(release_year). See methodology for the model + caveats.
This is a personal research project, not a service. Data refreshes monthly.