Tuesday, February 28, 2012

[csrjmslx] The most interesting graph

Consider a graph with nodes as people.  If two people have sex, draw an edge between them.  (Other variations, perhaps somewhat easier to obtain: who married whom, who biologically had children with whom.)

Assuming unlabeled nodes, what can be learned from this graph?

What node annotations (or even edge annotations) yield even more interesting insights?

I suspect a tremendous amount can be learned, but they may be things we wish we didn't know.  In choosing whom to mate, society's deepest prejudices become revealed: races, religions, other feuds we didn't even know existed. Have we made any progress?

No comments :