Thursday, August 14, 2008

One of the neatest graphs I've seen in a while

I think principal component analysis is one of the cooler applications of linear algebra (which I learned to analyze corporate bonds), but this is the coolest graph of principal components that I have ever seen. By simply using a matrix manipulation of a dataset based on the genomes of European citizens, you get a representation of the European genetic data based only on genetic data that looks remarkably like the geographical map.

See NYTimes article for more info.


No comments: