Inspired by Data Wrangling Westworld by Joel Carron.
Carron's original project built a dataset from episode scripts from springfieldspringfield.co.uk (which is now shut down). Additionally, the scripts were missing character names, so they needed to re-watch each episode and add the names manually.
Since this project is more about relationships than an analysis of the dialogue, I won't need to go that far.
Initial data is coming from the Westworld Wiki. It is either fetched through the API or scraped from the XML export when the API is insufficient.
Since no good source for all episode scripts is available, I may resort to analyzing subtitle files for character name mentions within a certain number of minutes, which may indicate that characters are talking to or about one another.
Subtitle files downloaded from http://www.tvsubtitles.net/tvshow-2081-3.html
Westworld really ought to have an API, like SWAPI or Marvel's API