[drone]

This happens more often than not on my trips through more rural areas (once, it sent me through a muddy corn field to avoid a traffic jam on Christmas night in rural Louisiana -- fortunately I understood what I was getting in to and had a vehicle equipped to handle it), and I've noticed it seems to have more to do with people who have google maps keeping their phones with them.

That is to say, it tracks where people drive, and when people drive in the same place often, it flags it as a common route and therefore must be a road! I have a hunting lease in rural Texas, and the most common trails within the property are all highlighted as roads in Google maps. In fact, as I cleared out a trail a year ago and started using that trail as a way to reach one point on the ranch, it has now also become highlighted!

You'd be a fool to cut someone's locked gate to take such a route, but it does offer them as routes to me were I to ask how, for example, to arrive at the entrance gate of the property north of us. The only legal (and passable) route is actually to go south, then east, then north, then west. However, google maps says "Hey, just take this trail straight north."

[notatoad]

>That is to say, it tracks where people drive, and when people drive in the same place often, it flags it as a common route and therefore must be a road!

do you know this to be true, or are you just guessing?

[jerf]

Well, between "Google is tracking people with cell phones to look at their common patterns and doing AI on the data", "Google actually sent out a surveyor to a hunting lease to track this guy's new trail", or "Google purchased someone else's survey of the new hunting trail", I find the first one most plausible by quite a bit.

I live in a less rural area but I've still noticed a handful of things in my area marked as more complete roads than they should be. Nobody in person would mistake them for even the lowest grade of official road in use around here.

[notatoad]

I find "google purchased a set of map data that said this creek bed was a road" to be much more plausible than any other scenario

[cowsandmilk]

Google regularly sends surveys to people showing they are tracking their movements. Like you’ll eat at a restaurant and get a survey about it. They don’t even hide this fact that they do this:

https://support.google.com/surveys/answer/6315313?hl=en

They ask you questions about train rides and bus rides as well. Google doesn’t buy data any more, they collect it.

Why wouldn’t they be tracking where people drive and building maps about new roads from that?

[notatoad]

>Google doesn’t buy data any more

again, is this something you know to be true, or are you just guessing?

[drone]

I know it to be true. Outside of the textual example I posted in my original post, here's a visual example: https://pasteboard.co/IloR6Hv.png

The highlighted tracks are private, mud-ridden trails on private property. If I'm sitting in that house, and ask Google to direct to me to the address directly north of my location, it will direct me to take that trail to the north. Which, BTW, will lead me to a locked gate and another trail on the other side of the fence. The correct answer is to send me south to the county road, east to the next county road, north to the next county road, and then west via county road due north of where I am. Those trails are not roads that google would otherwise know about.

[Somasis]

I've also had this happen to me- driving from Boone, North Carolina to Charlotte, I usually avoid highways because the interstates are a bit overwhelming to me. I took one different turn as I was trying to avoid downtown in a rural city I was driving through; they were having a block party and the usual route was blocked off. It's routed me through that turn ever since, so saying Google is learning from where people drive even enough to start making routes, that doesn't seem too farfetched for me.