|
Cellular forums Home > Archive > Garmin GPS > January 2006 > mapping story on npr this morning
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
mapping story on npr this morning
|
|
|
|
| PonderosaSports.com 2006-01-10, 5:49 pm |
| NAVTEQ, Google Maps, TeleAtlas - Are they frauds? Are they knowingly
providing false data? Is NPR planning a follow up?
Our business, Ponderosa Sports, is off by over 3 miles. The Idaho
state reference at the Idaho Department of Transportation in Boise is
off over 500'.
| |
| Jack Erbes 2006-01-10, 5:49 pm |
| PonderosaSports.com wrote:
> NAVTEQ, Google Maps, TeleAtlas - Are they frauds? Are they knowingly
> providing false data? Is NPR planning a follow up?
>
> Our business, Ponderosa Sports, is off by over 3 miles. The Idaho
> state reference at the Idaho Department of Transportation in Boise is
> off over 500'.
I don't think those are frauds, I think they present bad data in good
faith.
NavTeq's data goes back to the map data in USGS topographic products
and/or the TIGER map data from the U.S. Census Bureau. I don't have a
clear understanding of the merging or migration of that data, maybe you
can figure it out here:
http://www.census.gov/geo/www/maps/CP_OnLineMapping.htm
The autorouting maps are vector maps based on the USGS, TIGER, or
whatever data. NavTeq can correct or update their vector maps as they
want but they can't fix the errors in the original data.
Every USGS topo map I have ever seen has the road I live on mis located
by something on the order of 1/4 mile or so south of its actual location
and only partially depicted (modern extensions not shown). And every
version of every mapping software I have ever seen had, until just
recently, that same road mis located to the exact same position. So I
conclude that all of the companies started with the same data.
Microsoft Streets & Trips 2006, which credits both NavTeq and TeleAtlas,
finally shows that road in the correct position and, in addition has
extended and added to the road to accurately reflect its current length
and configuration. A new error added is that one of two new roads now
shown at the top of the road is misnamed.
If anyone is interested in this the road is:
Altons Avenue, Ellsworth, Maine, USA 04605
The road starts at 44-29-37N 68-25-11W (read degrees, minutes, seconds).
I think some of the major sellers of mapping software have websites
where people can submit info or suggestions for updates and corrections
to map products. I suppose it takes time to verify and implement the
changes.
The situation in the NPR story is not unique, I'll bet there are many
thousands of situations like that.
Jack
--
Jack Erbes in Ellsworth, Maine, USA - jackerbes at adelphia dot net
(also receiving email at jacker at midmaine.com)
| |
| Hans-Georg Michna 2006-01-10, 5:49 pm |
| On Tue, 10 Jan 2006 11:38:38 -0500, Jack Erbes wrote:
>The situation in the NPR story is not unique, I'll bet there are many
>thousands of situations like that.
Jack, all,
in my experience map errors are very widespread. When I drive I
encounter them all the time.
Two underlying problems are:
1. The map making companies are very few, hardly more than two
(NavTeq and Teleatlas), so we have little competition, high
prices, and low quality.
2. The method of creating a fixed map, then selling it until an
expensive update is created a year later at the earliest, is
very questionable. I'd propose to gather and process user tracks
instead and have a process in place to update the maps
near-instantly.
Hans-Georg
--
No mail, please.
| |
| Phil Wheeler 2006-01-10, 5:49 pm |
| Hans-Georg Michna wrote:
> On Tue, 10 Jan 2006 11:38:38 -0500, Jack Erbes wrote:
>
>
>
>
> Jack, all,
>
> in my experience map errors are very widespread. When I drive I
> encounter them all the time.
>
> Two underlying problems are:
>
> 1. The map making companies are very few, hardly more than two
> (NavTeq and Teleatlas), so we have little competition, high
> prices, and low quality.
>
> 2. The method of creating a fixed map, then selling it until an
> expensive update is created a year later at the earliest, is
> very questionable. I'd propose to gather and process user tracks
> instead and have a process in place to update the maps
> near-instantly.
>
Thinking about it, it's surprising we have maps as good as we have.
When I travel in more remote parts of the world, getting even a half-way
decent map can be a challenge. Pretty maps, yes .. but often there are
major errors.
And when I hike and think about the quality of the USGS topo maps (and I
seldom find errors in those not related to lack of update), I marvel at
the amount of work that had to be done to get the data to build thosed
maps -- for the most part before there was satellite data available (and
most of which did not rely on aerial data). The entire survey process
for some very remote and hostile areas is pretty impressive!
Phil
| |
| 4MLA1FN 2006-01-10, 5:49 pm |
| On Tue, 10 Jan 2006 18:26:43 GMT, Phil Wheeler <w6tuh-ng7@yahoo.com>
wrote:
>Thinking about it, it's surprising we have maps as good as we have.
that's funny phil. i was just about to post that same comment. for
me, the glass is more than half full and it'll only get better. i find
the directions i get to be quite accurate/usable except for one or two
memorable failures.
like others, my house is off by about 1.5 blocks. except for national
highways, it seems the map companies should be getting their data from
the local governments, who have to keep accurate records. maybe they
do.
| |
| David Lee 2006-01-10, 11:48 pm |
| 4MLA1FN wrote...
> like others, my house is off by about 1.5 blocks. except for national
> highways, it seems the map companies should be getting their data from
> the local governments, who have to keep accurate records. maybe they
> do.
Does it put your house on the right road? A lot of (maybe even most) house
numbers are interpolated between a very few geo-referenced properties. In
the UK it's generally not so bad since our roads are usually fairly short.
I imagine it could be very much more of a problem where house numbers run
into thousands!
David
| |
| Dana DeLouis 2006-01-10, 11:48 pm |
| For fun, I was testing my Garmin City Select map with the street mention in
the article...
About the middle of the street...
N34.09557 W118.25738
--
Dana
"4MLA1FN" <eff1fan@yahoo.com> wrote in message
news:rof7s1tppv5ot4g
30dj9cvr4j7e4p6f3tu@
4ax.com...
> http://www.npr.org/templates/story/...storyId=5140674
>
> in case you missed it, it's a short story about online navteq-based
> mapping sites and one place where they can't get you.
| |
| 4MLA1FN 2006-01-11, 2:48 am |
| On Tue, 10 Jan 2006 21:51:42 -0000, "David Lee"
< davidlee_malvern@don
t.use.this.bit.hotmail.com> wrote:
>Does it put your house on the right road? A lot of (maybe even most) house
>numbers are interpolated between a very few geo-referenced properties.
hi dave. yes it is on the correct street. i recently realized that
it interpolating. our street goes from 3000 to 3200. however, in
noticed in city navigator, it ends my street with 3300. that explains
why it's off as much as it is. should be an easy fix. :)
| |
| Hans-Georg Michna 2006-01-11, 5:48 pm |
| On Tue, 10 Jan 2006 18:26:43 GMT, Phil Wheeler wrote:
>Thinking about it, it's surprising we have maps as good as we have.
>
>When I travel in more remote parts of the world, getting even a half-way
>decent map can be a challenge. Pretty maps, yes .. but often there are
>major errors.
>
>And when I hike and think about the quality of the USGS topo maps (and I
>seldom find errors in those not related to lack of update), I marvel at
>the amount of work that had to be done to get the data to build thosed
>maps -- for the most part before there was satellite data available (and
>most of which did not rely on aerial data). The entire survey process
>for some very remote and hostile areas is pretty impressive!
Phil,
I agree, but the survey process is also unnecessary for many
purposes. Collecing user tracks would be a much more intelligent
way to create excellent maps, for two reasons.
1. The maps are accurate particularly where people actually
move.
2. The maps would reflect where people can actually move, for
example, not against one-way traffic or not across a narrow
ravine.
Hans-Georg
--
No mail, please.
| |
|
| Dana DeLouis <delouis@bellsouth.net> wrote:
>"4MLA1FN" <eff1fan@yahoo.com> wrote:
>
>For fun, I was testing my Garmin City Select map with the street mention in
>the article...
>About the middle of the street...
>
>N34.09557 W118.25738
Interestingly enough, Google Local also gives no indication
that's anything but a through street.
--
Joel Plutchak "Prescriptive lexicographers enjoy being grumpy.
plutchak@{VERYWARM}mail.com They spend a lot of time denouncing words and
explaining what the rules used to be."
- Richard Bready, Encarta
| |
| Jack Erbes 2006-01-11, 5:48 pm |
| Hans-Georg Michna wrote:
> Phil,
>
> I agree, but the survey process is also unnecessary for many
> purposes. Collecing user tracks would be a much more intelligent
> way to create excellent maps, for two reasons.
>
> 1. The maps are accurate particularly where people actually
> move.
>
> 2. The maps would reflect where people can actually move, for
> example, not against one-way traffic or not across a narrow
> ravine.
>
It is a good idea if it can be done. It would probably have to be
voluntary and initiated by the user to avoid implications of invasions
of privacy.
Getting the track info explained and submitted might be unwieldy.
Microsoft Streets & Trips 2006 has a feature that lets me click on any
spot on the map and send info to to M$ about anything on the map that I
want to comment on, including missing or incorrect map features. And
that happens in a brainless manner when I am online. The only question
is, is anyone reading the mail and what will they do with it?
For the Garmin and Magellan packages, a similar process for submitting
the same kinds of info would require a little more effort on the part of
the user.
But I could see where if they saw that thousands of people were driving
somewhere where there was no road it might prompt them to investigate
that. And then they could delete all the roads that no one uses because
they might not really be there? :> )
Jack
--
Jack Erbes in Ellsworth, Maine, USA - jackerbes at adelphia dot net
(also receiving email at jacker at midmaine.com)
| |
| Phil Wheeler 2006-01-11, 5:48 pm |
| Hans-Georg Michna wrote:
> On Tue, 10 Jan 2006 18:26:43 GMT, Phil Wheeler wrote:
>
>
>
>
> Phil,
>
> I agree, but the survey process is also unnecessary for many
> purposes. Collecing user tracks would be a much more intelligent
> way to create excellent maps, for two reasons.
>
> 1. The maps are accurate particularly where people actually
> move.
>
> 2. The maps would reflect where people can actually move, for
> example, not against one-way traffic or not across a narrow
> ravine.
>
Surely you jest. If topographic maps were done that way (assuming user
tracks could be taken as accurately documented data), we would have very
good info on mountain trails but not much on the area off the trails ..
since most hikers are on trail and those who do serious cross-country
are not going to be carring a turned on GPSR while they scramble.
Phil
| |
| Phil Wheeler 2006-01-11, 5:48 pm |
| Jack Erbes wrote:
> But I could see where if they saw that thousands of people were driving
> somewhere where there was no road it might prompt them to investigate
> that. And then they could delete all the roads that no one uses because
> they might not really be there? :> )
>
>
A new concept: Mapping by poll!
| |
| 4MLA1FN 2006-01-11, 5:48 pm |
| On Wed, 11 Jan 2006 11:27:27 +0100, Hans-Georg Michna
<hans- georgNoEmailPlease@m
ichna.com> wrote:
>Collecing user tracks would be a much more intelligent
>way to create excellent maps, for two reasons.
this would just give you where roads are located (ignoring the problem
of tracks from off-roaders). the problem is see is getting the road
names. is the user to annotate the track with street names?
| |
| 4MLA1FN 2006-01-11, 5:48 pm |
| On Wed, 11 Jan 2006 16:36:13 GMT, Phil Wheeler <w6tuh-ng7@yahoo.com>
wrote:
>A new concept: Mapping by poll!
and it's corollary: mega traffic congestion :) (if your map doesn't
have infrequently driven roads, you'll never find those "secret"
alterantive routes around traffic.)
| |
| Hans-Georg Michna 2006-01-12, 5:48 am |
| On Wed, 11 Jan 2006 16:35:08 GMT, Phil Wheeler wrote:
>Hans-Georg Michna wrote:
[color=darkred]
>Surely you jest. If topographic maps were done that way (assuming user
>tracks could be taken as accurately documented data), we would have very
>good info on mountain trails but not much on the area off the trails ..
>since most hikers are on trail and those who do serious cross-country
>are not going to be carring a turned on GPSR while they scramble.
Phil,
I think all available data should be used for mapping. User
tracks are a particularly interesting part, but obviously don
t always provide complete coverage.
Hans-Georg
--
No mail, please.
| |
| Hans-Georg Michna 2006-01-12, 5:48 am |
| On Wed, 11 Jan 2006 12:00:20 -0500, 4MLA1FN wrote:
>On Wed, 11 Jan 2006 11:27:27 +0100, Hans-Georg Michna
><hans- georgNoEmailPlease@m
ichna.com> wrote:
[color=darkred]
>this would just give you where roads are located (ignoring the problem
>of tracks from off-roaders). the problem is see is getting the road
>names. is the user to annotate the track with street names?
There are various ways to solve this problem. Asking the user is
one. The old method that's been used so far is another.
Hans-Georg
--
No mail, please.
|
|
|
|
|