[The New York Times Magazine, 10 December 2006]
By RICHARD MORGAN
Not long ago, Dirk Brockmann, a theoretical physicist, was chatting with his friend Dennis Derryberry, a cabinetmaker, about his interest in human travel patterns -- the types of paths that people trace as they move around by foot, by car, by train, by plane. The problem, Brockmann explained, was how to find data about such voyages. Derryberry suggested Brockmann take a look at Where's George, a popular Web site that tracks the location of U.S. bills. As a result, this year, in a January issue of the journal Nature, a new-and-improved science of human travel was born.
In the era before modern travel, when people did move about, they did so in a relatively limited manner. Nowadays, the patterns of human travel are presumably far more varied and erratic, but until the paper by Brockmann (which he published with two colleagues), scientists hadn't been able to generate a reliable model of just how varied and erratic.
The creators of Where's George weren't paying attention to human beings, but they unwittingly amassed valuable data about the millions of journeys that we make from point to point. Here's how the site works: you select a bill from your wallet and enter the denomination, series and serial number, as well as your ZIP code, into the Web site, which registers the date of entry. The more you do this -- and the more others do this with the same bills at later dates -- the better the record of where the bills (and by rough proxy, their various spenders) have traveled. The site has millions of users, and Brockmann's paper analyzed 1,033,095 reports submitted on the dispersal of 464,670 bills.
In the end, the authors were able to create a model that allowed them to predict the probability of a bill staying within a 10-kilometer radius over a period of time -- as compared with drifting 100 or 1,000 kilometers over the same period. This was more than a mere academic accomplishment: in order to understand, say, the spread of avian flu, you need to understand the patterns of mobility by which the disease will likely be spread.
back to home page