Collect screenplay data for either nlp or character networks for films / TV series

Screenplay

There are no comprehensive online databases of screenplays. There are some projects of particular series or films with acceptable data formats we can access, like Sex and the City or Star wars original trilogy.

Besides those, some websites try to collect screenplays. The most popular one is IMSDB. At first glance, some factors make it harder to use these scripts. Like different formats, formatting, and script versions, but it’s workable.

Closed Caption

It is much easier to access the caption of films and series. We could scrape websites storing subtitles (e. g. YIFY). However, they are not official, so maybe it would be better (and cooler) to use sources with a more official look.

I checked three streaming services Netflix, Amazon Prime and HBO GO. Netflix’s and Amazon’s captions seem to be relatively easily accessible. However, further examination is needed.

Developer tools -> Network -> Caption file. Amazon stores its caption in a Timed Text Markup Language 2 (TTML2) format. You only need to find that file. Netflix starts the name of the caption file with a “?o=”.