The data (u|dys)topia

Take one simple conversation with a neighbor, add a few late night walks in the dark, and it develops into a complete utopian/dystopian vision of the not-so-distant future.

First, the conversation: we were talking about everyday marital strife, and she mentioned how her husband wants to redo the roof himself — it’s a relatively straightforward project, he’s retired and has the time, and they’d save a lot of money. Her response is that his life cannot possibly be worth the money saved. Result: intractable argument. How can either side prevail? She spoke about couples taking turns “winning” arguments as a way out.

My thought is that it’s actually a question of data. What’s the actual risk that he’d injure or kill himself? If he takes the proper precautions (like a safety line, ladder spotter, etc.), can he make working on the roof safer than driving the car to the store? If yes, her argument goes away. If not — particularly if amateur roof work is dangerous even if the homeowner believes they’re taking precautions — then she wins.

Of course, it’s a little more complicated than that. We don’t operate on pure reason (or even primarily reason, really).

Extreme example: we know plane travel is safe. If you took a commercial flight every day for the next 18 thousand years, that would still only bring your odds of being in a crash to about fifty-fifty. Yes, it’s that improbable. But even knowing that just doesn’t completely do it on a blustery winter takeoff, and some people are always going to be simmering with barely-suppressed panic every time they’re strapped into one of those heavy, huge metal tubes packed with fat tourists as it lumbers down the runway and hefts itself unnaturally into the air.

So even if the roofing accident statistics are on his side and the actual risk is low, she might worry herself sick anyway… and that has an emotional cost that needs to be factored in.

But let’s get back to the data question.

Are these statistics even available? Does anyone collect data on how many amateur roofing projects ended in an accident, and (importantly) what safety measures were being taken at the time? Something could possibly be cobbled together on accidental death statistics, since I know there are collections of those — but I doubt the details we really want are out there.

Here’s where it gets interesting. It wouldn’t be all that big a step to get all the data I want, and oh so much more.
Suppose we each had a few minor, low-tech implants — like a tiny chip in each hand & foot, and a central implant that would collect data from those chips and transmit it (even via existing cell networks) back to a central location. All we need to transmit is a steady stream of data on the orientation in 3D space of those chips and (a bit less precisely) the person as a whole.

It’s incredible what we could get out of that. Respective location of hands/feet/torso translate to body position, and movement patterns (you can deduce velocity of movement based on change in data points) that would be fairly easily translated into walking, running, jumping, driving a car, sitting, sleeping, walking on a flat surface, climbing a ladder, carrying heavy items (very easy on soft ground, but it would also change your walking pattern even walking on concrete), walking on an incline, falling….

Even with no reliance on existing geo data, collections of human movement would easily define buildings, offices, bedrooms, bathrooms, baseline ground level, roads, walking paths, etc..

With no video whatsoever (and no complicated analysis of images or video) a medical program could notify you if you were favoring your right leg recently when climbing stairs — even though you hadn’t noticed — and at the office you sit at something of a slant. Useful, no?

Through normal data mining techniques we’d be able to know how many people fell from the roof of their own home while hammering there, for example.

More abstractly, we could also gather concrete statistics on accidents and deaths, and patterns of movements leading to those misfortunes, and (one more feature for that implant…) notify you when you matched a high-risk pattern. With access to multiple streams of data, you can even widen the net to dangerous interactions, like warning you when a drunk driver going far faster than surrounding cars is going to be overtaking you on the highway in a minute.

And obviously, law enforcement could have already been notified when the guy swayed and stumbled out to his car, after downing 94 gulps of very-likely-beer in a bar. Heck, they might already be there, if patterns of a bar fight were picked up earlier.

This is all very possible with current technology. Actually, leveraging current technology (and other available data) would open entire new vistas that I haven’t even thought through yet… I’m still agog at what we could do with just this little stream of data.

Take all of the buzz that’s already been brewing for a while about applications of the GPS tracking that’s very likely already built in to your mobile phone, and multiply it by crazy. “We know where he is, what’s he doing?”

And of course, this is where the dystopian aspect swings in. How could privacy issues realistically be covered? …because how much money would that information be worth? And is the minutely-observed life properly worth living? Would we become paranoid about accidental mimicry of really-bad patterns? (Honey, the police are here… apparently brushing the dog hairs off your coat looked a little too “stabby”…). Could the data be used as court evidence? Could it be tampered with? What kind of nastiness could a random disgruntled employee with access to the data pull off?

Food for thought, anyway; it’s time for another nighttime walk for me.