text2image [teddavis.org] is one of those "data art" works that balances on the line of total uselessness and visual fascination. While there is no immediate purpose, one still spends minutes playing around with it.
It is an online tool which allows the user to transform text of any sort into a "visualization" of that very content. The results are deterministic, in that they are consistent to the given input, but extremely dynamic through its variance. Notably, "the application of these images is a complete unknown."
Well, I know a use: as a game to try to figure out how what the heck is the algorithm that turns the characters into a visual representation. As long as the word "infosthetics" comes out quite beautifully, I am satisfied.