Feb 22

Brady Forrest

Brady Forrest

Progress Always Needs Help From Users

Exton, a town in the UK, has begun putting up signs to prevent trucks from getting stuck on a tight road. Why are they getting stuck? Bad mapping data. In an article from The Mail On Sunday Exton resident, Brian Thorpe-Tracey says:

'About two years ago we noticed a real increase in drivers using the lane. Vehicles are getting stuck and having to reverse back up, damaging the wall and fence. There's even a piece of metal embedded 12ft up in a tree which looks like it's come off a lorry.
'When I've asked drivers why they are using the lane they say they are just following satnav.

This is a case where user contributed data would save the day. That isn't always the case. Sites like Open Street Map rely almost entirely on auto-submitted user-contributions are well-equipped to handle this. Teleatlas and NAVTEQ have begun accepting corrections via their Map Insight and Map Reporter (respectively). This is a good step on both of their parts, but the data still needs to be manually processed before it will be fixed (even then it can be months before partners receive it).

[via Techdirt]

tags: geo  | comments: 9   | Sphere It

Previous  |  Next

0 TrackBacks

TrackBack URL for this entry:

Comments: 9

  Jonathan Arnold [02.22.07 07:08 PM]

I have a similar problem at my house. For some reason, our little road isn't connected to the major road just down the hill on any map computer, including the nav in our car. So when I email directions, or tell someone on the phone, I have to be sure to warn them to ignore the directions once they get close to my house, or else they go on a wild journey through the neighborhood.

  Matt Turner [02.23.07 02:31 AM]

If i'm not mistaken, this is the same road a coach went up full of passengers (elderly people if i remember correctly) and got stuck. Eventually it had to be driven cross country to escape (minus the passengers)!

If it's not the same road it was a very similar circumstance where the driver was following sat nav and got stuck.

  Paul Martin [02.23.07 09:36 AM]

One would think that headline treatment like this would be sufficient to get the attention of the mapping services. Apparently, the problem of maintenance and error correction aren't very high on their list. We have similar problems in my state; there is massive progress in road improvements, and sleepy little two lane roads are turning into limited access four lane roads, but routes suggested by mapping services ignore the new data.

This is a problem crying out for the harnessing of collective intelligence. If Google or Yahoo or Mapquest were to start doing this, they would grab market share very quickly.

  peter renshaw [02.24.07 04:25 AM]

'... Bad mapping data ...' ~ the bane of users & cartographers.

This is something that is carefully thought about and avoided in the creation of traditional topographic maps through careful checking for 'ground truth'. The quality of the data is paramount.

"Is this junction on the map the same as what is on the ground ?".

Here's the catch. Printed maps take a long time compile & correct & can be out of date when published. While GIS (Geographic Information Systems) coupled with GPS allow reems of data to be generated fast but with little thought to accuracy.

So until companies producing mapping products using GIS information "ground truth" their data, problems such as the ones described will continue.

  Rich Lyon [02.26.07 04:43 AM]

There is a place in Yorkshire, UK (sadly, and to the detriment of my story, I don't know where) in which drivers are directed to cross a ford which is about 2 feet deep. While this is entirely feasible for the horses and carts for whom the base map was first constructed, it's deep enough to stall a sub-compact. Jolly amusing.

  wabanstar [02.26.07 08:00 AM]

People think there data should either already be mapped or will be soon, for what reason? If they had any inkling of the vast amounts of data that has never been mapped, may never be mapped, was mapped badly originally, and the amount of data that changes yearly in a large dataset and the costs associated they would be amazed that they can even use their sat navigation unit at all. Basically if you are in major urban corridor perhaps expect 75% to 98% accuracy, if not the quality could vary wildly.

  Michael [03.27.07 09:54 PM]

Maps are generally not to be trusted. Getting away from the high tech end of things, I do a lot of tramping here in New Zealand, both on and off track. The position of some features and tracks can be out UP TO ONE KILOMETER. This can be... problematic in cliff prone country and we lose the odd person or two each year. These are the govt created official topomaps mind you, so most providers of maps use their data rather than say walking every mile of terrain themselves with hand help gps units...

The map is NEVER the terrotory.

  W3C-Webdesigner [11.26.07 05:50 AM]

Google accepts corrections made by users, too.

But all companies have the same problem with the Web 2.0 hype: If they start involving the users, there will be always users, who manipulate it to their own needs. What will be the solution for that?

  brady [11.26.07 09:28 AM]

Now Google accepts updates from users. At the time that this article was written they did not.

You are correct we will have to watch for Map Spam and other similar issues. However, systems can be trained to watch for lots of edits at once and for contentious edits. Overall I believe that the error rate will decrease with user input.

Post A Comment:

 (please be patient, comments may take awhile to post)

Type the characters you see in the picture above.