3 ms·

I trusted Google Maps untill few weeks ago it sent us down a road that became undriveable and dangerous -- costing us about two hrs. Although, this did happen

by dk8996 11y ago

I trusted Google Maps untill few weeks ago it sent us down a road that became undriveable and dangerous -- costing us about two hrs. Although, this did happen in Martinique, for the most part I've had good experiencea with Google Maps in US.

ghaff 11y ago
In the Western US, there are a lot of roads that range between unpaved but well-graded and generally fine for passenger car use to serious high-clearance 4WD in good weather only. In my experience, Google Maps doesn't routinely send you down seriously questionable roads. But, if you drive around those areas, it behooves you to really understand the range of roads out there and be properly cautious rather than blindly depending on GPS--or, for that matter, a map. The biggest problem is really that a lot of people come from areas where road is 99% equivalent to passable by passenger car under most to all weather conditions and seasons. That's just not the case in many places, e.g. Death Valley.
- maxerickson 11y ago
  Rural roads in the United States are something to be especially wary of when using OpenStreetMap derived maps. They are often un-updated TIGER data from 2008, prior to Census doing a big push on accuracy. There are issues with roads being poorly located (or not existing), but there are also issues with little or no differentiation between roads, so a nice new asphalt highway might be classed the same as a barely passable dirt road. One reasonable strategy would be to stick to things that are labeled as state or US highways. (with enough scale, it should be possible to use traffic data to figure out which roads should be prioritized, but programs like OSMAnd are not yet doing that)
  - ghaff 11y ago
    >with enough scale, it should be possible to use traffic data to figure out which roads should be prioritized It wouldn't be perfect but, yeah, I'd expect to see some correlation between traffic and road condition--at least for a given area. There's just a lot of nuance in road condition though. People who take their Jeeps off the beaten track every weekend in national forests and parks have a pretty different set of standards from an urbanite who is renting a car with a doughnut spare and no other special gear.
    - maxerickson 11y ago
      The goal with OSM is certainly to sensibly classify roads and even to collect information like the surface type and width. I guess with that part of the comment I was partly wondering if other data providers have done similar things, taking some vaguely useful source data and trying to enhance it using automated methods, leading to the poor navigation talked about in various threads here.
- crygin 11y ago
  One of my favorite examples of this is using Google Maps to find your way from Monument Valley to Natural Bridges or vice versa -- sends you right up/down the Moki Dugway, which despite being part of a state highway is not a road for everyone.

by dk8996 11y ago

ghaff 11y ago

In the Western US, there are a lot of roads that range between unpaved but well-graded and generally fine for passenger car use to serious high-clearance 4WD in good weather only. In my experience, Google Maps doesn't routinely send you down seriously questionable roads. But, if you drive around those areas, it behooves you to really understand the range of roads out there and be properly cautious rather than blindly depending on GPS--or, for that matter, a map. The biggest problem is really that a lot of people come from areas where road is 99% equivalent to passable by passenger car under most to all weather conditions and seasons. That's just not the case in many places, e.g. Death Valley.

maxerickson 11y ago
Rural roads in the United States are something to be especially wary of when using OpenStreetMap derived maps. They are often un-updated TIGER data from 2008, prior to Census doing a big push on accuracy. There are issues with roads being poorly located (or not existing), but there are also issues with little or no differentiation between roads, so a nice new asphalt highway might be classed the same as a barely passable dirt road. One reasonable strategy would be to stick to things that are labeled as state or US highways. (with enough scale, it should be possible to use traffic data to figure out which roads should be prioritized, but programs like OSMAnd are not yet doing that)
- ghaff 11y ago
  >with enough scale, it should be possible to use traffic data to figure out which roads should be prioritized It wouldn't be perfect but, yeah, I'd expect to see some correlation between traffic and road condition--at least for a given area. There's just a lot of nuance in road condition though. People who take their Jeeps off the beaten track every weekend in national forests and parks have a pretty different set of standards from an urbanite who is renting a car with a doughnut spare and no other special gear.
  - maxerickson 11y ago
    The goal with OSM is certainly to sensibly classify roads and even to collect information like the surface type and width. I guess with that part of the comment I was partly wondering if other data providers have done similar things, taking some vaguely useful source data and trying to enhance it using automated methods, leading to the poor navigation talked about in various threads here.
crygin 11y ago
One of my favorite examples of this is using Google Maps to find your way from Monument Valley to Natural Bridges or vice versa -- sends you right up/down the Moki Dugway, which despite being part of a state highway is not a road for everyone.