Unboxing Letterboxd: An Exploratory Data Analysis of Movies, Genres, Reviews, and Audience Perception of the Female Gaze in Women-Directed Films
About the Project
Letterboxd began as a “Goodreads for film”, a social cataloguing platform where cinephiles could maintain personal film diaries. Over time, however, in the age of curation under platform capitalism (Kompatsiaris, 2024), it evolved into a performative online space for signalling cultural capital. Simply sharing movie reviews, as was common through blogs and forums, is no longer enough. Users log every movie-watching experience, engage with others’ reviews, and curate recommendation lists. While IMDB had long served as a film tracker, Letterboxd breathed new life into digital cinephilia, transforming it into a curatorial practice deeply entangled in the algorithmic sociotechnical infrastructures of contemporary platforms. From an audience reception perspective, the Letterboxd dataset is a goldmine for examining film trends, fan discourses, and review-writing practices, offering a window into how movie watching is not just an aesthetic experience but a quantifiable, curatorial hobby.
Developed for the Scripting Languages course taught by Tim van de Cruys as part of the Advanced Master’s in Digital Humanities at KU Leuven for the academic year 2025-26, this project explores the Letterboxd dataset through exploratory data analysis and preliminary text analysis techniques. Sourced from Hugging Face and likely scraped, the dataset is used here strictly for educational purposes, serving as a basis for examining patterns in contemporary film culture through user-generated data.
The project combines exploratory analysis of structured Letterboxd metadata with an initial exploration of movie review text. After basic preprocessing and dataset overview, the analysis focuses on identifying trends in genres, ratings, and release years, as well as patterns in genre co-occurrence in contemporary films. It then turns to a subset of reviews from films selected from Rotten Tomatoes’ list of the best 21st-century films directed by women. Using this corpus, a preliminary text analysis is conducted to identify recurring semantic themes and patterns in audience discourse, with a view to understanding how these films are received in relation to perceptions of the director’s female gaze.