A if a picture is worth a thousand words, how much can a thousand pictures tell us? The problem when looking at high-dimensional data, or trying to fit models to them is that our visual system just is not made for “getting” more than two dimensions in static images. There are many approaches to circumvent this problem, most notably interaction plots (wiki), but in the end the most simple solution is to use depth from motion, one of the most powerful 3 cue we have.
While generating an unlimited number of 3D-graphs is a breeze - here you go. The final frontier for me was to combine these graphics into one moving animation. This is extremly easy to do in Imagemagic, one of my favorite open-source graphics programs. All functions are accessed via a command-line interface. So that that can be used to generate text from images, or combining many different pictures into a single animated gif. That means that you have to make up new tasks for your student assisstant. If you have a bunch of jpgs in your folder the following command combines them into an animated file.
> > convert -delay 10 *.jpg animated.gif > >
The delay argument works as expected to control the offset between pictures.
The website is no looker, but in fact the useage-section gives an great number of extremly useful examples from working with text, to animations, or blurring.