There are a couple simple pieces to making this work:
1) The individual data points are drawn with the “scatter”, where x, y are obviously the coordinates and s is the size of the dot. I also used an alpha value to have some transparency to accommodate overlapping data, for example:
pyplot.scatter([x],[y], s= [z], color = c, alpha = 0.4)
Note I was plotting one point at a time, hence the [x] etc instead of just ‘x’ if your data was in an array already.
2) In order to get a matrix like layout, with (0,0) in the upper-left, I reversed the y values and labels when plotting them. I also set the axis range just outside the data that was plotted so it looked cleaner.

Due to the data I was using and the form I processed it in the code may be a little funky for your purposes, but I cleaned it up to a relatively small demo, which I’m sure you could adapt. Here is some imaginary data of Rock-Paper-Scissors versus my husband and the full code example for plotting Rock-Paper-Scissors Matrix:

Enjoy!

Zia

Could you please describe what commands you used to get the plot? It looks really cool, but I can’t seem to find anything similar in the gallery (http://matplotlib.org/gallery.html).

Cheers,
Michael