Fun with Mathematica, Wikipedia, and Graphviz
One of the fine gentlemen at work showed me a project his son is working on in school. InPhO (http://inpho.cogs.indiana.edu/) analyses the Stanford Encyclopedia of Philosophy to build an ontology. There is some human input, and some of the areas are not fleshed out, but it's pretty groovy and they're off to a great start.
Since reading up on philosophers and their influences struck a chord with me, I cursed my colleague that he had condemned me to wasting the rest of the day clicking around his son's website. I hit one of my favorites, Baruch Spinoza, and noticed that the set of influences and those influenced wasn't heavily populated. Since writing stupid programs and throwing them into smarter programs is something along the lines of a way of life for me, I hacked together a little program to crawl Wikipedia from a known starting point, and crawl afferent and efferent influence declarations until I reached leaf or ingenerate nodes. One particularly vexing issue with Wikipedia is inconsistent naming for the same person (Hegel was a particularly painful example) and I leaned on a little help from my old friend Levenshtein to combine some names, but some duplicates made it through. In addition, some concepts appeared to be first-class citizens, and while I was able to flag some, others snuck through.
With the data in hand, I was able to ask Important Questions, such as,"What is the chain of influences connecting Aristotle to Tom Waits?"
How about Plato and Pynchon?
I bet you can't connect Ovid and Chomsky!
OK, that's fun enough, but I wanted pictures! I generated a plot in Mathematica, but it was the least fun plot ever. Even in 3D.
I emitted a dot file for Graphviz, and it promptly created the second-to-least fun graph ever. Growing tired of the exercise, I imported the dot file into OmniGraffle Pro and crashed it about a hundred dozen times, making incremental tweaks to the top ten influencers (i.e., the huge vertices) and the layout to make the behemoth image that may or may not be above, depending on whether Posterous wanted any part of this business. If you have a plotter handy, I highly recommend seeing if it will become sentient upon putting this mess to paper.
I poked around the plotted graph to see if it was worth the trouble to which I had gone, and found some amusing little connections, giving a lens into the minds of Wikipedia editors and philosophers.
I think perhaps a LITTLE more influence is called for there. Then again, it appears founders of religions are under-appreciated.
I'm not going to comment on the fact that there is no line connecting the upper oval to the lower one, other than to say that it's obvious Wikipedia is a wretched hive of scum and infidels.
One thing that surprised me, that I alluded to above, is how many non-philosophers were drawn into the graph. A lot of authors, and even some directors (The Wachowski brothers and J. J. Abrams, for example) made appearances.
And to top it all off, even Jimmy Buffett snuck in there, Hal Holbrook and Michael Crichton in tow. Fortunately, it appears they have influenced even fewer people than Jesus or Muhammad.
If it weren't already 3:20am, I'd probably try to refine the data collection a little more. However, two hours spent on this is enough. Most of it was spent tweaking in OmniGraffle, but I couldn't sleep and thought I might as well do something useless.
Comments [0]