Hacker News new | ask | show | jobs
by turingfeel 1610 days ago
I think this is because Google compiles these cast lists based on search proximity prevalence. So if you look up some actor and a movie in the same query many times it starts to assume the actor is in that movie. Maybe they combine this with some base scraping but I think it's highly unlikely they use some API or hand craft the lists.
2 comments

I always assumed it was scraped from IMDb or TMDB. It usually includes both character and actor names together, which seems hard to glean from search queries alone. Ordering them based on search volume could make sense though.
Definitely possible. I think they got a lot of backlash from scraping a music lyric site without permission so they might be more hesitant to do so in other areas unless they have clear cut permission.
So if there were two actors that were commonly confused, likely they'd end up showing as if they were both in the same movie.

Sounds like a great way to get inaccurate results.