304,713 utterances and 220,579 conversational exchanges extracted from 617 movie scripts involving 9,035 characters. The collection integrates movie-level metadata like IMDB ratings and genres with character-specific attributes including gender and billing position.
Use Cases
- Train natural language generation models using the conversational exchanges and utterance text to mimic cinematic dialogue.
- Conduct sociolinguistic studies by correlating the gender character metadata with linguistic patterns in the utterances.
- Predict movie popularity or critical reception by analyzing dialogue features against the IMDB rating and IMDB votes columns.
- Investigate character hierarchy and social dynamics by comparing the credit position metadata with the volume of conversational exchanges.
Strengths
- 304,713 individual utterances mapped across 10,292 character pairs.
- Character metadata includes gender for 3,774 individuals and movie credit position for 3,321 individuals.
- Movie-level attributes include release year, genres, IMDB rating, and number of IMDB votes.
- Structured as 220,579 conversational exchanges derived from 617 distinct films.