Movies and TV shows with plots, cast and ratings dataset cover

Mostly Right

Updated27 Oct 2026

Movies and TV shows with plots, cast and ratings

Name: subtitle_chunks
Creator: Mostly Right
Published: 2026-11-06T20:56:25.039Z

movies: Data · Columns
subtitle_chunks: Data · Columns

Updated 21 Oct 2026

2.1 GB across 15 parts, newest sealed 21 October 2026 at 15:40 UTC.

The first build wrote 15 parts and 12.2M rows.

ONE_ROW_PER_CHUNK passed.

id

integer

TMDB id of the film, matching the movies table's id; empty when the film's imdb_id is not in the movies table.

No distribution published

Valid: —
Missing: —
Distinct: —

imdb_id

string

IMDb identifier of the title, such as tt0133093; join to the movies table's imdb_id.

No distribution published

Valid: —
Missing: —
Distinct: —

lang

string

ISO 639-1 language of the subtitle text; en.

No distribution published

Valid: —
Missing: —
Distinct: —

subtitle_id

string

OpenSubtitles subtitle file identifier the chunk was cut from; one file per title.

No distribution published

Valid: —
Missing: —
Distinct: —

year_dir

integer

Release year under which the OPUS OpenSubtitles v2024 corpus files the film; empty when the corpus folder is not a year.

No distribution published

Valid: —
Missing: —
Distinct: —

chunk_index

integer

Zero-based position of the chunk within the title's subtitle file.

No distribution published

Valid: —
Missing: —
Distinct: —

start_seconds

decimal

Start of the chunk's first line, seconds from the start of the film.

No distribution published

Valid: —
Missing: —
Distinct: —

end_seconds

decimal

End of the chunk's last line, seconds from the start of the film.

No distribution published

Valid: —
Missing: —
Distinct: —

start_time

string

start_seconds as hh:mm:ss.

No distribution published

Valid: —
Missing: —
Distinct: —

end_time

string

end_seconds as hh:mm:ss.

No distribution published

Valid: —
Missing: —
Distinct: —

text

string

Dialogue of the chunk, about 30 to 60 seconds and at most about 2,000 characters, with sound-effect brackets and speaker labels removed.

No distribution published

Valid: —
Missing: —
Distinct: —

n_lines

integer

Number of subtitle lines joined into the chunk.

No distribution published

Valid: —
Missing: —
Distinct: —

hi_ratio

decimal

Share of lines in the chosen file that looked hearing-impaired (bracketed sound or speaker label) before cleaning; 0 to 1.

No distribution published

Valid: —
Missing: —
Distinct: —