I think you could approximate this for movies using a social graph - I had a hobby project when I got back into programming a few years ago which attempted to generate recommendations based on the director's and actors' "creator" social graphs. It worked surprisingly well for such a hack.
Sounds interesting. How would one build it for 'influence'? E.g. I would imagine it to be vast and utterly fascinating to map the influence graph of Tarantino :)
I am no expert, but extracting data related to the word influence from IMDB for Tarantino (or any actor/director) is quite doable. Even by employing rudimentary NLP techniques using NLTK and Freebase/wordnet APIs I can gather all Tarantino related documents in which the word influence(or its many synonyms) have appeared. Then I can look for proper nouns in the vicinity of the word 'influence' which will give me a list of all people who have influenced Tarantino.
Doing the next level analysis is hard, finding the degree of influence, building a timeline graph of how Tarantino was influenced over a period of time etc.
I find the topic of extracting such graphs quite fascinating. I was thinking of doing the something similar for Prof.Snape in the Harry Potter series.
