Screw the details, all I wanted to do was play with this fun new toy, so I searched for my own papers. The programme assigns facial features to various aspects of the publication - for example, the age of the face corresponds to the age of the paper, and bigger smiles mean more citations compared to the expected number. The gender and race are usually assigned randomly, as in the examples below, although you can standardise them so that the other features stand out more.
Here's my best paper (38 actual citations compared to 16 expected - it was in PNAS!)
And here's my worst (only cited 6 times when 13 citations were expected)
Hey, that last one reminds me of something. What could it possibly be?
Oh, now I remember.
(Explanation at Mad Hatter's place).
I wonder if we could use the same code to assign faces to our Euro 2008 pool results. Okham will be looking a little happier in today's version.
But not much.