Seattle

I didn't know about intersections, but knew that some retail stores and businesses actually do this. Devices around the stores will listen for Bluetooth transmissions or even WiFi signals from phones to see where people are congregating, how long they're in the store, and all kinds of things. It's why Android and Apple changed to use random addresses for phones.

You and your devices give off a lot of unexpected signals that can be monitored and tracked. Some things could be good like monitoring dwell time at a business or even queueing time at Disneyland, others could be used for marketing at a company. Just something to consider the next time you use WiFi at a business.

Not the intersection itself exactly, but city gear along the way. Bluetooth was used for comms for the first field wireless meter reads because all field communication was shoddy back then. I know it sounds stupid now but it was amazing vs the old way of each meter having to be actually physically read. There were employees whose whole job was to drive a city truck around a specific route to collect data, which was seen as a huge productivity upgrade from getting out at each place and looking at the meter. As much gear as possible was deployed with bluetooth connectivity so they could do drive by "remote" reads. Houston standardized on it for awhile before more modern techniques for reads came out. Water meters, sensors for public works water main, well levels for drains under overpasses, stuff like that all over the city. City gear is absolutely everywhere, we're all just conditioned to ignore it. You couldn't write to the device from the field, but if you polled it it would answer with a reading for whatever it was measuring.

At some point they realized some people left their bluetooth on on their phones (which wasn't the case when the initial deployment happened, bluetooth was seen by most as a battery sucking crap technology) and by comparing the bluetooth ping logs at two points they could approximate driving speeds to a decently accurate degree. You couldn't use the data to pinpoint a specific user really and you couldn't pinpoint speed exactly so it was no use to law enforcement, but it was fabulous data to model traffic on.