I scraped 3,060 Letterboxd user comments and ratings for Dead Poets Society, then uses R to build a word co-occurrence network colored by Ridge regression coefficients to visualize which words and themes are associated with higher or lower ratings.
The Data
I downloaded 3,060 user comments (and their star ratings) for Dead Poets Society from Letterboxd, the social platform for film lovers. After filtering for English-language comments and removing common stopwords, I was left with a rich vocabulary of the words people actually use when talking about this film.
The Method
The analysis has two layers:
Word co-occurrence network. Using the quanteda text-analysis package in R, I built a Feature Co-occurrence Matrix — essentially counting how often words appear together in the same comment. The 1,250 strongest connections were kept and visualised as a network graph, where linked words tend to show up in the same reviews.
Rating prediction. I then fitted a Ridge regression model (glmnet) to predict a comment’s star rating from the words it contains. Each word gets a coefficient: positive means the word is associated with higher ratings (shown in green), negative means lower ratings (red). Neutral words sit in a pale yellow middle ground.
What Stands Out
The emotional core is positive. The densest part of the network — words like beautiful, amazing, masterpiece, poetry, passion, dreams — glows green. People who reach for these words tend to rate the film highly. This cluster captures the movie’s central themes of seizing the day and the beauty of self-expression.
Character names anchor the conversation. Neil, Todd, Anderson, Keating, and Mr. Williams all sit near the centre, tightly connected to everything else — they’re the vocabulary everyone shares regardless of opinion.
Negative words live on the edges. Red-tinted nodes like sad, apart, shame, and favourite (which, interestingly, co-occurs with more critical commentary) drift toward the periphery. The word stupid also appears in an orange hue, often showing up in comments that are more mixed or critical.
A “why it matters” cluster on the right. Words like beauty, passion, human, write, read, art, engineering, medicine, law, business form a distinctive green cluster — reflecting comments that discuss the film’s central tension between pursuing the arts and conforming to practical career expectations.
Takeaway
Most people love this film. The network is overwhelmingly green, and the most connected, central vocabulary is deeply positive and emotional. Critical voices exist but are quieter and more scattered — the few red nodes at the margins of an otherwise glowing map of admiration.
Since this is not the first time I analyze moviecomments the rather clean scripts can be found here: https://github.com/thhaase/letterboxed-dead-poets-society
more letterboxd analysis: