Support NeoGAF

[RiccochetJ]

The author of this article decided to take the "Six Degrees of Kevin Bacon" and downloaded the entire IMDB data set to figure out who would be the most connected person in film if you exclude porn. He does actually look at porn as well in the article.

But is Kevin Bacon the ‘most connected’ or “most important” person in the film industry? To answer this question, I downloaded the entire IMDB database. This is a rich dataset (like over 3.6 million people in over 5.1 million titles dating back to 1894 rich) and once you start “extracting the network” from this data it gets huge. By “extracting the network” data I just mean that we are interested in not just the people and the movies but also all of their connections. To make this analysis easier, I used only movies made after 1960 with at least 4,171 IMDB ratings (the mean) and an overall rating of above 1.0. Even still, once we extract the network, we get over 12,000 movies, over 50,000 people, and half a million unique connections!

Closeness centrality is defined as the sum of a node’s distances to all other nodes and can be used to evaluate which node is most important for disseminating information. In other words, a node with high closeness centrality is more ‘embedded’ in the network than other nodes.

The image below shows a subset of connected nodes from the network. The size of the nodes corresponds to their centralities for the four different measures, and the colors correspond to their different “communities”

Let’s apply these centrality measures to the entire network of 50,000 people. The ten people with the highest degree centrality are:

1. Robert De Niro (actor)
2. Nicolas Cage (actor)
3. Jerry Goldsmith (composer)
4. Samuel L. Jackson (actor)
5. James Newton Howard (composer)
6. Scott Rudin (producer)
7. Bruce Willis (actor)
8. Mark Isham (composer)
9. Michael Caine (actor)
10. James Horner (composer)

This means that these men have worked with the most people during their careers. While they may not all be household names, this list really isn’t that surprising. All of these men have made many movies with many people for many years. Robert De Niro has been a principal in 82 movies with 613 unique collaborators.¹ Jerry Goldsmith passed away in 2004 but was also a principal in 82 films. The ever-prolific Nick Cage has made 73 films with 602 and people, five of them last year alone.

It's an interesting article to see how you can look at something as fairly trivial as the IMDB database and use that data to tell a story. There are a lot of different graphs in the article as they manipulate the data to understand different metrics. I find it absolutely fascinating!

For example, here's where they bring in William Shakespear into connections:

 

[Rentahamster]

and an overall rating of above 1.0.

That's why Kevin Bacon's not on the top.

 

[triplestation]

caralho!