What did the Google Streetview Car Collect?

[Driver8]

@Rhythmdvl

Right, but isn’t that information inherently transient? Today it’s “Rhythmdvl Coffee Wifi,” tomorrow I replace my router and now it’s “Rhythmdvl**’s** Coffee Wifi.” Or I sell the business to a flea circus and they don’t change those particular hardware settings. It wouldn’t take many changes in a city to degrade the original information to low-quaility, particularly if Google has no way to discriminate between good and bad data.

In aggregate, it probably isn’t that transient. People do occasionally change their WiFi profiles but not every day. I doubt the system would be structured to ask “If we see Rhythmdvl Coffee Wifi we are on Main Street”. It would rather be “Hmm, we see 9 WiFi hot spots here, 4 of which match exactly (down to the NAT address) what we stored when we drove Main Street a few months ago. There is a good chance the machine is on Main Street”. The information could also be combined with other sources of location information, like cellular towers in the case of mobile devices, to provide an estimate of location that while not guaranteed to be accurate, may be good enough in most cases.

[CurtC]

@Rhythmdvl

Right, but isn’t that information inherently transient? Today it’s “Rhythmdvl Coffee Wifi,” tomorrow I replace my router and now it’s “Rhythmdvl**’s** Coffee Wifi.”

I believe they log the MAC address (instead of the SSID) for that very reason. Although MAC addresses can be spoofed, it’s very uncommon to change the MAC address of your router.

[Reply]

@Driver8

In aggregate, it probably isn’t that transient. People do occasionally change their WiFi profiles but not every day. I doubt the system would be structured to ask “If we see Rhythmdvl Coffee Wifi we are on Main Street”. It would rather be “Hmm, we see 9 WiFi hot spots here, 4 of which match exactly (down to the NAT address) what we stored when we drove Main Street a few months ago. There is a good chance the machine is on Main Street”. The information could also be combined with other sources of location information, like cellular towers in the case of mobile devices, to provide an estimate of location that while not guaranteed to be accurate, may be good enough in most cases.

Good explanation.

[_Undecided_Adrian]

The slip-up was apparently discovered by Google Germany because there was an inquiry in their Wifi data mining when it became publicly known that such data was gathered. Privacy is a much bigger concern in Germany than in the US, and the minister of consumer protection for example is rather critical of Google.
I personally still think their intentions are good, and use their services extensively, but with this and the Google Buzz launch they could have handled things better.