MassHighway Traffic Cameras Reveal Google Map Errors (Long)
Bob Malme
access many of its traffic cameras in the Greater Boston area
including along I-93, I-95(128), and US 3. While one page shows all
the camera images at one time. If you check on them individually they
give you the live feed, images denoting what direction you're looking
at, and a Google Map of the location. In going through all the cameras
and checking on the Google maps, I noticed many errors. These include:
Route 3A south of Boston continuing north of Neponset Circle onto
Morrissey Blvd. and Old Colony Ave. in South Boston; MA 30 Continuing
along US 20 to Kenmore Square (and in some more detailed maps,
beyond); MA 3A north of Boston continuing along US 3 south of I-95
(128) all the way to the Alewife Brook Parkway (MA 16). Both US 3 and
MA 3 are shown on Comm Ave east of the BU Bridge; and MA 6A on the
Cape not ending at US 6 at the Sagamore bridge but continuing south of
the Cape Cod Canal to MA 28 at the Bourne Bridge. Extending the maps
beyond the web page limits also leads to the discovery of further
errors such as US 44 ending at MA Exit 7, not continuing along Route 3
to Exit 8 and ending at MA 3A and MA 228 not ending at GW Blvd in Hull
but following Nantasket Avenue northward to the tip of Hull.
Some of these errors could be attributed to the way the Mass. Highway
Inventory is organized. It lists by route the town it is in, the road
it's on, where the road goes in its entirety in the town (regardless
of whether the state route stays on it) and then breaks down the route
into smaller segments and lengths for where the route does go. It is
the third category that may confuse some map coordinate errors. For
example for MA 30 in Boston, it is listed as running on Commonwealth
Avenue. Comm Ave is first listed as running from the Newton line to
Arlington Street, then the segments involving MA 30 are noted which
indicates it ends at US 20 which then continues on Comm Ave until
ending at MA 2 in Kenmore Square. Google editors got the ending of US
20 correct, but not MA 30. (The 2007 Mass State Map which still the
most recent online shows MA 30 continuing with Route 2 to Arlington
St. in the Boston inset, so MassHighway's own map readers can't follow
the inventory very well either it appears.) These may account for many
of the errors, but not all. MassHighway also considers alternate
routes as being co-routed with their parents, though unsigned, that
could account for the extension of Route 3A along US 3, but not the
extension of 3A into South Boston along Morrissey Blvd., which was
never part of 3A or Route 3 for that matter.
Given the amount of errors found in one metropolitan area, I would
have a hard time trusting Google Maps the next time I need directions.
(Do any others, like Mapquest, have the same errors?) I know someone
who works for the firm that supplies Google it's data. If there are
any other errors that you think Google should be aware of, I can
collect a list and forward it to my contact, to see what happens.
Having bad maps on a page that supposed to help travelers doesn't
reflect well on MassHighway, regardless if they're not responsible.
larry_scholnick
I much prefer Bing Maps (maps.bing.com) over Google Maps. Bing's
Bird's Eye view isn't quite as ambitious as Google's Street View, but
it is far easier to use.
Jeff Morrison
> Given the amount of errors found in one metropolitan area, I would
> have a hard time trusting Google Maps the next time I need directions.
> (Do any others, like Mapquest, have the same errors?) I know someone
> who works for the firm that supplies Google it's data. If there are
> any other errors that you think Google should be aware of, I can
> collect a list and forward it to my contact, to see what happens.
> Having bad maps on a page that supposed to help travelers doesn't
> reflect well on MassHighway, regardless if they're not responsible.
When Google switched data providers, its error rate went up
exponentially. Many US shields are incorrectly rendered as state
shields. There are scores, if not hundreds, of state highways in just
Iowa and Wisconsin that have been truncated or off the books
completely for six, ten, or 29 years and are still appearing, while
recent changes are halfheartedly marked. Google Maps has stopped
differentiating paved Iowa county roads from gravel ones and often
incorrectly names them. Not to mention that every lettered route in
Missouri now appears to be duplexed with Route Y. (I have no idea how
these mistakes affect point-by-point directions.)
Quite frankly, there are so many errors, and it's so systemic, that I
believe a simple list is nearly impossible. Because it's a dynamic
system, I don't think I can just say "change the shield here" and
"remove the shield there" and "this should be a more/less prominent
line". Even if I could, the list would be very, very long and hard to
describe without lots of additional maps and screenshots.
If any course of action can be suggested, it would be "Look at the
state DOT maps and adjust markings accordingly. For Pete's sake, IA 73
was killed before I was born."
Paul D. DeRocco
>
> When Google switched data providers, its error rate went up
> exponentially. Many US shields are incorrectly rendered as state
> shields. There are scores, if not hundreds, of state highways in just
> Iowa and Wisconsin that have been truncated or off the books
> completely for six, ten, or 29 years and are still appearing, while
> recent changes are halfheartedly marked. Google Maps has stopped
> differentiating paved Iowa county roads from gravel ones and often
> incorrectly names them. Not to mention that every lettered route in
> Missouri now appears to be duplexed with Route Y. (I have no idea how
> these mistakes affect point-by-point directions.)
I expect that most of the errors are in the numbering and naming, not the
actual location and connectivity. Given the increasing popularity of GPS and
turn-by-turn, the whole idea of route numbers is becoming less important.
But some of the errors are amusing. I noticed in Barstow, CA, that the old
alignment of CA-58 north of the town, which used to be US-466, and which is
now not a numbered route at all, has shields listing it as both US-58 and
CA-466.
--
Ciao, Paul D. DeRocco
Paul mailto:pder...@ix.netcom.com
Otto Yamamoto
> Quite frankly, there are so many errors, and it's so systemic, that I
> believe a simple list is nearly impossible. Because it's a dynamic
> system, I don't think I can just say "change the shield here" and
> "remove the shield there" and "this should be a more/less prominent
> line". Even if I could, the list would be very, very long and hard to
> describe without lots of additional maps and screenshots.
Google has never been good, but since switching to Tele-Atlas, they've
seriously tanked. US 9W's numerous demotions to NY 9W nearly matches the
number of error shields you find in the field, never mind the fact that
they show NY 218 running all the way to Saugerties(it runs from Highland
Falls north to Cornwall-old US 9W with a short concurrency with US 9W at
West Point).
Microsoft may not be so great at OS's(Vista is a misery), but they are
moving streets ahead in both online and software based mapping. Streets
and Trips is the best consumer level software out there from what I can
see(I'm going to buy that and Delorme and spend some time with them to
get a better idea of the quality).
--
Otto Yamamoto 'Eat flaming death, fascist media pigs!'
D. Kirkpatrick
<f13a16e6-0a2d-414a...@j19g2000vbp.googlegroups.com>,
Bob Malme <rma...@verizon.net> wrote:
> Given the amount of errors found in one metropolitan area, I would
> have a hard time trusting Google Maps the next time I need directions.
Part of the problem is that this is driven mostly by GPS and longitude
and latitude. If the L&L and GPS are off on a day that the snapshot
is taken, then the map remains broken. Ditto if the 911 street
address database is off.
I have found similar problems myself.
I wanted to offer people directions to a building near me for a
meeting. When I inserted the actual street address it gave a push-pin
a decent 1/3 of a mile elsewhere.
When I inserted the wrong but close address it came within feet of
what I wanted.
I had to send the link with a caveat that while the address was
different on the map, that it was for purposes of making the Google
Map show the right place. This allowed both the correct area map but
also a "bird's eye" view of the building.
I've also found this to be a problem in very rural areas. When
attempting to map an address near the NY state line in a rural area, I
kept being told that the address was non-existant. Yet I was using
the actual street address that is required for 911 response service.
before street numbering was necessary for 911 mapping, this was a
rural route designated area.
Again, I had to fudge an address just to get the map to come up close.
I'd not venture comment on the other services but like most they are
also L&L and GPS driven.
Paul D. DeRocco
>
> Microsoft may not be so great at OS's(Vista is a misery), but they are
> moving streets ahead in both online and software based mapping. Streets
> and Trips is the best consumer level software out there from what I can
> see(I'm going to buy that and Delorme and spend some time with them to
> get a better idea of the quality).
My experience with MSS&T vs. SAUSA is that the former makes really legible
maps, but is a bare-bones program with only minimal features, while the
latter is super-configurable and extensible, but makes butt-ugly maps. I
wind up using the former a lot and the latter almost never.
I still use DeLorme's Topo USA sometimes while I'm on the road, but it has
the same ugly map look, and its 3D rendering makes my machine feel like an
old PC XT compared to the zippy display from Google Earth. They really need
to throw out their software base and start over.
Marc Fannin
> I wanted to offer people directions to a building near me for a
> meeting. When I inserted the actual street address it gave a push-pin
> a decent 1/3 of a mile elsewhere.
>
> When I inserted the wrong but close address it came within feet of
> what I wanted.
>
> I had to send the link with a caveat that while the address was
> different on the map, that it was for purposes of making the Google
> Map show the right place. This allowed both the correct area map but
> also a "bird's eye" view of the building.
My advice, FWIW: Use coordinates in this context. They're always
embedded in the URL that you get when you take the "Link" link near
the upper-right corner, then you can apply them to the URL:
http://maps.google.com/maps?ll=[lat/long separated by comma]; tack on
a
&z=[1 through 19]
if you want the zoom to be different than the default. Granted, this
takes away the pushpin and its description window, but it doesn't let
there be two different addresses floating around out there, one of
which might be mistaken for the correct one.
_________________________________________________________________________
Marc Fannin|musxf579 @hotmail.com|http://roadfan.com/ (m.t.r FAQ, etc.)
Garrett Wollman
Otto Yamamoto <ros...@yamamoto.cc> wrote:
>and Trips is the best consumer level software out there from what I can
>see(I'm going to buy that and Delorme and spend some time with them to
>get a better idea of the quality).
The recent versions of Street Atlas are awful. Older versions have a
much more usable user interface, but of course their map databases are
woefully out-of-date.
-GAWollman
--
Garrett A. Wollman | What intellectual phenomenon can be older, or more oft
wol...@bimajority.org| repeated, than the story of a large research program
Opinions not shared by| that impaled itself upon a false central assumption
my employers. | accepted by all practitioners? - S.J. Gould, 1993
Otto Yamamoto
> My experience with MSS&T vs. SAUSA is that the former makes really
> legible maps, but is a bare-bones program with only minimal features,
> while the latter is super-configurable and extensible, but makes
> butt-ugly maps. I wind up using the former a lot and the latter almost
> never.
Yeah, S&T isn't the fanciest programme out there, but unlike with it's
OS's, Microsoft found something that works and has stuck to it.
Paul Anderson
> Google has never been good, but since switching to Tele-Atlas, they've
> seriously tanked.
When did they switch? I've already logged three errors to Tele
Atlas--all within two miles of my house, in a semi-rural town! Some of
these errors would cause people to get seriously lost.
Paul
Otto Yamamoto
> When did they switch? I've already logged three errors to Tele
> Atlas--all within two miles of my house, in a semi-rural town! Some of
> these errors would cause people to get seriously lost.
Just this year. Not sure of exactly when.
Bob Malme
As of the end of last week (10/9), Google Maps dropped TeleAtlas, at
least from the text at the bottom of the maps. They now read Google
instead, and all the maps have been revised. Most of the errors I
listed for MA have been fixed on the new maps (or old, could be these
are the maps they used before contracting with TeleAtlas). The one
sort of annoying change is they have changed bannered routes
(alternate routes like 3A) from route symbols to text running along
the streets, along with the street names. This makes for crowded text
at certain zoom levels. They also have placed these bannered routes
along their parent routes, as they officially run according to the
MassHighway Route Inventory. Thus Memorial Drive is signed in places
as US 3, MA 3 and 'Massachusetts Route 3A.'
David Z Maze
> As of the end of last week (10/9), Google Maps dropped TeleAtlas, at
> least from the text at the bottom of the maps. They now read Google
> instead, and all the maps have been revised. [...] The one
> sort of annoying change is they have changed bannered routes
> (alternate routes like 3A) from route symbols to text running along
> the streets, along with the street names. This makes for crowded text
> at certain zoom levels. They also have placed these bannered routes
> along their parent routes, as they officially run according to the
> MassHighway Route Inventory. Thus Memorial Drive is signed in places
> as US 3, MA 3 and 'Massachusetts Route 3A.'
...or more egregiously, at what I believe is zoom level "11" (on my
screen, about the right size to see all of 128 but no more) the
Southeast Expressway is principally US-1, but also MA-3, and is named
"Massachusetts 3A"; I-93 appears to skip around downtown Boston and
Quincy. If I zoom in once more I can see the I-93 shield and the
"Southeast Expy" label.
There's a lot of nitpicky errors like that now that I'm looking around,
including US-3 extending past MIT, and the mysterious "N" route numbers
popping up in various places.
(I do, however, appreciate "Metropolitan Storage Warehouse, Fire Proof"
being explicitly labelled as a building in Cambridge. :-)
--dzm
Jimmy
> On Sep 27, 11:48 pm, Otto Yamamoto <ros...@yamamoto.cc> wrote:
>
> > On Sun, 27 Sep 2009 21:23:30 -0400, Paul Anderson wrote:
> > > When did they switch? I've already logged three errors to Tele
> > > Atlas--all within two miles of my house, in a semi-rural town! Some of
> > > these errors would cause people to get seriously lost.
>
> > Just this year. Not sure of exactly when.
>
> > --
> > Otto Yamamoto 'Eat flaming death, fascist media pigs!'
>
> As of the end of last week (10/9), Google Maps dropped TeleAtlas, at
> least from the text at the bottom of the maps. They now read Google
> instead, and all the maps have been revised. Most of the errors I
> listed for MA have been fixed on the new maps (or old, could be these
> are the maps they used before contracting with TeleAtlas).
Another plus: Google Maps now shows property lines for some
Massachusetts towns. This makes it very easy to spot things like
paper streets and former rail lines. It's also interesting to see how
lot sizes vary in different towns.
I took a quick look: Boston, Cambridge, Belmont, Brookline, Chelsea,
Concord, Newton, Waltham, Lexington, Quincy, and Woburn have the
property lines, while Somerville, Arlington, Watertown, Milton,
Everett, Lynn, and Winthrop don't.
Jimmy
Garrett Wollman
David Z Maze <dm...@mit.edu> wrote:
>(I do, however, appreciate "Metropolitan Storage Warehouse, Fire Proof"
>being explicitly labelled as a building in Cambridge. :-)
Do they have Kaplan Furniture Company as well? (Is that hoarding even
still visible on the building any more?) How about the NECCO Factory
(that isn't any more)?
John S
switched to TeleAtlas under a five year deal. TeleAtlas would also
receive corrections from Google users. NAVTEQ had been purchased by
Nokia and TeleAtlas is owned by TomTom. There have been a lot of
complaints about TeleAtlas map quality in the USA.
Google is now reporting, and TeleAtlas confirms that Google has dropped
TeleAtlas for USA maps. Over the past few days, the TeleAtlas copyright
would come and go depending on zoom and panning, but it appears to be
gone now.
As Google has been building its own map capabilities, I'm curious why
the TeleAtlas agreement was as long as 5 years. I'm also curious what
the breakup fees were. Perhaps there none, which answers my first
question as well.
See
http://searchengineland.com/no-more-teleatlas-google-goes-it-alone-for-maps-data-27584
(Google also now lists Route 3 in Woburn and Arlington, MA as also being
3A, which does not exist in those towns. Under TeleAtlas, it was listed
as only 3A--also wrong.) Route 3A north of Boston only exists where the
non-limited access US 3 has been replaced by the freeway US 3, from
Burlington to the New Hampshire state line.