Perhaps you, like I, have dutifully copied a serial number from a one dollar bill bearing a red stamp into the database at wheresgeorge.com. And perhaps you, like I, thought as you did so: "Surely nothing can ever come of this. How many people are actually as fastidious as I am, saving one particular bill in their wallet until they are reunited with the Internet?"
It turns out we were wrong. Something did come of it.
It's clear that this video was made by the scientists -- they don't shy away from using the jargon as a matter of principle. But it works. Because we're seeing the algorithms run as the narrator strings together words we don't understand, we're still able to get the essence of what they did. The video also has a strong voice and sense of humor, which sets it apart from many of the other, more educational winners.
Wow, that was heck of cool. I want to understand that grouping algorithm better.
One question I had is how the spatial dependence of desire to log bills on the internet (high in SF, low in Appalachia, say) might impact the results. To first order I would expect that it would just depress and enhance the number of connections to a place, and not so much where those connections go to and come from, which I guess is in some sense 'linear' to the networking results. But then you wonder if, say, Philly and Pittsburgh are a little more connected because bills don't get logged in Appalachia, even if they stop there.
Fascinating data, and great graphics. I am a sucker for anything that shows Montecarlo.