GMaps Goes From Being Correct, to Being Wrong
Written by Chad on September 22nd, 2008Some may recall my earlier post about the differences between GMaps and Google Earth (How GE is wrong and GMaps is correct on the naming of a lot of roads). Well, with Google latest change, switching to TeleAtlas from Navteq has now made that post null and void, as now both GMaps and Google Earth are both WRONG now (and in looking around the local area.. there are a lot of side roads and back roads that are now completely wrong in their names).
So, what does that mean? Maybe nothing at all. Or maybe if I want to give someone directions, I will have to hand draw them a map because of they try Google.. they will get names of roads that do not match what they will see when driving.
I wonder if there has ever been a major comparison of road-to-road between TeleAtlas and Navteq, be interesting to see just what else is wrong between the two, and which one is more correct.
22
AM
If I didn’t know better I’d say you were trolling with that title.
Everyone knows that TeleAtlas is better than Navteq.
[Translate]
22
AM
Knows how it is better how though? Has someone compared them in great detail yet? (Would love to see that data if it has been done)
I look at it this way, I am giving someone directions to a certain place.. and if they look it up on GMaps.. they may get roads that are named wrong. So now they are driving around looking for RoadA when it is really named RoadB.
So, in a sense, the post title is right.. GMaps use to have roads around me named correctly.. now they are not named correctly and the information IS wrong.
If OSM had a good way to get directions, I would point people to that instead.
[Translate]
22
AM
I was being difficult.
We choose Navteq or TeleAtlas based on the local situation. Some cities seem to be better with one or the other. I prefer TeleAtlas for the Phoenix area given their quality over Navteq’s. It is a zero sum loss/gain from an accuracy view because for every problem you note, I can show you where it is better now.
Personally I don’t blame Google for dropping Navteq given Nokia owns them now. You can’t be funding a competitor’s R&D.
[Translate]
22
AM
Both are wrong, and both are right.
Depending on the area you are working on looking at.
Mapperz has compared both together and there are fundamental differences.
But remember that these datasets are used and designed for SatNavs. There is a long way for those datasets become accurate and fully suitable for cartographic products
one or two errors?
http://maps.google.co.uk/maps?f=q&hl=en&geocode=&q=no+name&sll=53.800651,-4.064941&sspn=8.857858,19.577637&ie=UTF8&ll=50.770638,-3.026133&spn=0.01851,0.038238&z=15&iwloc=addr
would you go on that railway?
no name is a place!
[Translate]
22
AM
You? Difficult? Never!
@Mapperz: No Name sounds good to me
But that sounds like a good reason why SatNavs seem to get people that follow them blindly into so much trouble.
[Translate]
22
PM
I for one welcome our new data overlords. TeleAtlas is far more name-accurate for my part of the world, primarily because they get data from the same company that services the local 911 centre.
[Translate]
23
AM
I agree with other posts; accuracy depends on location. I used a Garmin (NT) several years ago in Queens/Brooklyn NYC and data was awful. Many times an address would not even geocode due to the way Queens localities were organized. Then again, Queens is it’s own animal!
TA has made great strides in the more outlying areas in the last few years.
[Translate]
20
PM
Well instead of hand drawing, you could always fix the are on http://www.OpenStreetMap.org and give people that.
[Translate]
23
AM
Does anyone know how the road positional accuracy of TeleAtlas or Navteq compare to the USGS BTS road network?
In doing overlays of the BTS vector data on VE or GE maps you see substantial differences.
Thanks
[Translate]