Science: You Are Predictable
Image by AFP/Getty Images via Daylife
Using anonymous cell-phone records to determine people’s locations throughout the week, a physicist shows that he can predict a person’s location at any given time (within a couple miles) with 93% accuracy.
AOL News writes up the study:
Barabási first used himself as a test subject: He wore a GPS-enabled watch that repeatedly recorded his position from July 2007 to August 2008. As a popular academic, he attended conferences and meetings all over the world, yet most of the time he merely hopped between his home and his office. Even those long trips fell into a groove. “I was terribly boring,” he says.
The real surprises came when Barabási and his group began applying data-mining algorithms and advanced mathematics to a much larger set of people. A European mobile phone carrier — they won’t disclose which one — granted the scientists access to portions of the anonymous records of 50,000 mobile phone users, each of whom made, on average, at least one call every two hours. Chaoming Song, a co-author of the paper, says that because the data are made anonymous, the people the study tracked are effectively like particles in a gas that move and interact.
The carrier filed the location of the closest mobile tower whenever one of the individuals used the phone. From this, Barabási and his group could extract roughly — within a square mile or so — where a given phone and, by extension, its associated user were at a given time.
They found that most people stay close to home and, more intriguingly, that even the frequent travelers were no less predictable than the homebodies. Furthermore, they discovered that this phenomenon didn’t merely stem from the workweek — the fact that so many of us spend Monday to Friday in the same office. Weekend movements were no more random.
In effect, we are predictable even when we don’t have to be. A summary of the findings puts it this way: “Spontaneous individuals are largely absent from the population.”
We are creatures of routine. Though, this would probably get a lot less reliable if the grain of the data were finer.
Post Your Comment
You must be logged in to post a comment
T/S Members
Log in with your True/Slant account.
A point, and a story:
Point- this research actually confirms an observation made in the 1950’s by Guy DeBord, of the Situationists, that most people’s physical movements through a city are routine, limited, and entirely predictable. (To counter that narrowness, he prescribed something he called a “derieve”, a determinedly random walk through the city where one lives, in order to better explore and understand it.)
Story- once, taking an introductory college course in statistics, I saw a student stand up in his seat, scream, “I’M NOT A NUMBER! I’M NOT A NUMBER!”, and run headlong out of the lecture hall. There was a pause, and then the instructor, a pleasant woman who looked like Miss Marple, shrugged, and said, “On average, that happens once every other semester.”
[...] You Are Predictable Using anonymous cell-phone records to determine people’s locations – via TruesLant- throughout the week, a physicist shows that he can predict a person’s location at any given time (within a couple miles) with 93% accuracy. [...]
[...] know where you [...]