As our weekly readers know by now, I am a big fan of getting as much data as possible into a plot in a clear manner. David has done that in bubbleplot
. In a normal scatter plot he has managed to associate different series of data by use of color, and then get a third variable in there by changing the size of the marker. This same association and Z value could be done by having a three-dimensional line plot, but it would not be as clear as this. For some datasets this is the right visualization to use.
Bubbleplot was accomplished in MATLAB through the use of handle graphics properties. David wrote a wrapper, or small piece of code, around the normal PLOT calls to make it easy to change these property values.
I am also choosing this because it comes with a nice demonstration script that is clearly labeled. Brilliant code is nice, but making it easy for new people to try is even better.