Thursday, August 20, 2009

The Examined Life

I have two jobs, as I've mentioned before. The common thread in the two of them is to use software to hold up a mirror to one's life in some aspect or other, and learn what it really looks like. "What is normal?" is the big question I find myself attempting to answer.

Now, like most people, I can answer specific questions to describe myself: I am 5'10" tall, weigh *cough* pounds, have black hair (mostly black, anyway), brown eyes, and a goatee. I wear glasses to correct for nearsightedness and astigmatism. These are all physical properties. Let's call it a 0th-order description of myself.

These things change over time, in ways that can be measured. My hair is getting progressively whiter as time goes on. My height is more or less constant, my weight is, er, not. Let's call this a 1st-order description of myself. This is harder to measure, but can be done with some simple note-taking.

I could go further. My hair gets longer and shorter in cycles, according to my visits to the barber shop, for example. These cycles in turn may be lengthening as time goes on. This sort of thing is harder still to measure.

What about my behavior? Assume for the moment that the word "behavior" is well-defined, although it is not. I am sitting and using a web browser right now. Specifically, I am updating my blog, a 0th-order behavior. One could get a 1st-order behavior out of this by looking at my blog history, but this would at best be an approximation, under the assumption that all of my posts spring fully-formed out of my head and take only as long to write as they take to type, and that I do not throw away any half-written posts. (If only I threw away more, some of you complain...)

There are ways to capture this behavior a bit more exactly, and related behaviors besides. I could run network monitoring software of the type I write, which records open sessions between my laptop and the server and plots them according to time. I could do this for Kingdom of Loathing, the New York Times, and other sites that I like. (In fact, I have, and if anyone's interested I could post graphs) People trying to make changes in their lives frequently look at this sort of 1st-order behavior as an impetus for change: "I drink how much?", &c. Similarly, it's hard to indicate progress in a running or diet regimen without this sort of measurement.

So, why not a more holistic way of measuring oneself? I'm constantly amazed by how many people think the idea is somehow creepy. We all own mirrors, despite what we look like at 6am. But then, maybe we'd shy away from it if we had to use photographs instead of mirrors. (We'd probably also shy away from shaving) Me, I'm curious.

Other people and groups have been thinking along these lines too, by the way. There was an excellent interactive graphic in the NY Times recently, the result of minute-by-minute surveys of people. Fun to poke around with, I wish I had a graph like that for myself.


  1. I find massive data collection creepy mostly because if third parties get access to the data, they might do things with it that aren't in my best interest. Ignoring that risk, though, I think it's cool.

    On occasion, when navelgazing about how to optimize my life, I've found myself wanting one of those alarms that researchers use to randomly sample people's everyday behavior. Basically it's a beeper that goes off at random intervals, and when it goes off you're supposed to record data of some variety, whether it's your heart rate, your mood, or what it is you're working on. It's a little less holistic than what you're talking about, I think, but it could be a fair bit better than retrospective reporting about how on-task you were this week or what-have-you.

  2. Potential exposure to a third party contaminates the data, I think. For some things (monitoring a workplace network) that might be desirable. But most of these collection projects understand and want to avoid that sort of thing.

    I've seen those alarms -- one of the research groups I'm familiar with uses iPhones -- and they do seem like a neat idea. I've seen groups also try to use Twitter as a stand-in, albeit with some significant self-selection bias.