Wednesday, September 05, 2007


Super-hero social networks?

Analyzing social networks is all the rage these days. Even before things like Facebook and MySpace came around, social-network analysis was being used to determine subject-matter experts automatically, and to prototype email programs that would organize the mail by its likely importance to the recipient. Among other things, we're using it to help identify spam (more accurately, to help pick out the non-spam).

Well, in the “people with more time than sense” department, we now have a researcher who’s actually wasted spent his time analyzing comic-book fiction for social networks, and he’s reported his results:

Physicist Pablo Gleiser of the National Council for Scientific and Technical Research in Buenos Aires, Argentina, studied the social web within the fictional universe of Marvel comics, comprising 6486 characters in 12,942 issues. Taking two characters to be linked if they appeared in the same issue, he found a superficially realistic social network. A small fraction of characters — notably the superheroes themselves — had far more links than most others, acting as key social hubs. “The Marvel universe looks almost like a real social network,” says Gleiser (
OK, well. So the paper finds that the main characters in the fictional stories are the most richly connected in the long run. What a surprise.

And what an amazing waste of time.

Yes, but he’s just having fun, right? Hey, if you like to read comic books, why not throw in some statistical analysis of the characters and their relationships? You know, just for kicks!

Except that he claims to have government funding for this, grants from the Argentinian CONICET (National Council for Scientific and Technical Research) and ANPCyT (National Agency for Scientific and Technological Promotion), and the Italian ICTP (International Centre for Theoretical Physics).


This has to be a joke, and New Scientist was fooled or is playing along. I looked briefly, and I can’t find the cited grant numbers on the organizations’ web sites, and the author’s name only shows up on CONICET in his profile as an active physicist. The paper is dated 17 August. Maybe that’s the Argentinian version of April Fools’ Day.

One can hope.

1 comment:

Dr. Momentum said...

If it's really grant funded, I should think the funding source would require identifying which grant enabled the research. Unless this is some sort of proof of concept.

However, I can't imagine what the possible use is for a fictional social network.

A big TV universe crossover chart that I saw once was somewhat informative because it showed that crossovers were a lot more popular in TV than I expected.

The reason a Marvel Universe analysis is less interesting is that the innovation of the "Marvel Universe" is that it was one big common space that allowed superheroes to cross over into other titles. And they do it at the drop of a hat. In the early days of comics, titles were more like TV shows -- all with their own universes for the most part.

So it would seem to me to be neither useful nor entertaining.