The other day, travelling from an unfamiliar part of town back home through heavy traffic, I decided to just let Google Maps call the shots. I could have navigated myself, just using major routes that went towards areas I knew, but I decided to see what Google Maps did for me.
It took me on a rather interesting journey.
That’s not the most interesting bit, but let’s start with that.
It navigated me quite competently in the direction of home, using a route that I probably wouldn’t have chosen even if I’d been familiar with the area. That worked well. Then, in more familiar territory, I decided to let it have its head when it directed me off the main route and through a bunch of quiet back streets, before getting me back onto the main roads that I’d left.
It did that three times, and all in all, it seems those strange little detours saved me something like 20 minutes off the time I’d roughly expected to spend on the trip, bypassing road-works and heavy congestion. Anywhere that its live data was showing as being slower than alternate routes.
Then, a couple of days later, I am coming home quite late at night. I’ve got Google Maps running navigation, even though I don’t actually need it. Unexpectedly, Google Maps tells me to make a turn to go around an area that I would ordinarily go straight through. I’m curious as to why.
So, I decide to ignore it and continue on.
A couple of minutes later, I found out why. It was an RBT.
For those of you unfamiliar with Australian RBT setups, they’re “Random Breath Testing” by the police. They’ll set up on a piece of road (any day of the week, any hour of the day or night), where you cannot see them set up until you have no option to turn off or turn around. They pull vehicles over largely at random and test for blood-alcohol levels with an amazingly reliable electronic gadget, and sometimes also have saliva test-kits for a range of illegal drugs. Judging by the number of cars you’ll often find driverless and getting towed away, they’re generally pretty successful, even at times of day you’d not expect anyone to be drinking let alone drinking and driving.
They usually line up a half a dozen officers, pull people over until they’ve all got a driver to test, then let cars through until those drivers have been tested and cleared away, then pull over another batch. It’s really very efficient.
However, to Google Maps, it probably looks more like a traffic hazard. The drivers being tested are stopped for a couple of minutes, and all other traffic (in both directions) slows right down, because, well… uniformed police have that effect.
So, what Google Maps had spotted was traffic congestion in an odd spot, probably flagged it as some kind of hazard (accident or road-works), and looked around to see if it could save me time by routing me around “the problem area.” It could and it tried to. I just ignored it because I was curious as to why.
I’ve since looked at Google Maps at a couple of key locations and yes, you can pretty much always tell when there’s an RBT there just by looking at the Traffic Layer. Not that RBTs bother me. I don’t begrudge the extra couple minutes if I get pulled over (the saliva test can take a few minutes longer in cold weather, though, because the chemical reaction slows down), and of course there’s no other penalty for me, as I never drive under the influence of alcohol or drugs (and neither should you!).
It is a rather interesting, unintended consequence of having a lot of data, though.