Exploring Local
Mike Dobson of TeleMapics on Local Search and All Things Geospatial

News about the East Kilbride Expressway and Other Tidbits

December 18th, 2010 by MDob

Just some clean-up stuff today. I know that many of you are out there doing your holiday shopping and will have little time for anything as mundane as reading blogs. Me too. So happy holiday to all and —

Now on to something merrier.

‘Twas the night before Christmas and all though the house
not a creature was stirring not even a mouse.
The jolly old elf was preparing his route,
but Santa was a geographically ignorant old coot.
He looked in his atlas on the eve of that day,
but could not find the East Kilbride Expressway.

Man, I hate it when that happens to Santa. So, given that I had some free time, I thought I would help out. Very carefully, I entered the name East Kilbride Expressway into Google maps and that was where the trouble began. Apparently the people from Google Maps had read my last blog and were waiting to confound me on my next map search, not knowing I work working on Santa’s behalf.

The response to my query was this attempt at disambiguation. I had not known that there were than many East Kilbride Expressways in Scotland.

Disambiguation from Google

So, I began to page through the results. I mean it’s an expressway, right? Shouldn’t be hard to find. Are there a large number of East Kilbride Expressways in Scotland?. Hmmm. Google seems to think so and apparently its Street View vehicles have been on each and everyone one of them. The first one, I thought, was a likely candidate.

Good candidate

It looks like an expressway, doesn’t it?

Now for the East Kilbride Expressway in Dumfries and Galloway – this is an expressway?

Maybe it's a rural expressway?

Can’t be the right one. Next, let’s look at the East Kilbride Expressway in Falkirk

Oh, really, an expressway?

Not my idea of an expressway. Now the next one, well, it seems impossible that there is an East Kilbride Expressway, let alone any expressway in the Orkney Islands, but why not take a look?

Yowie

It seems to me that the cows are wider in the Street View Image than the East Kilbride Expressway shown here. And in this scene, along the same road, the driveway is wider than the East Klibride Expressway!

Maybe the driveway is the expressway - its wider

Let’s try Fife

NOT an expressway

Guess not.

Actually the first example I showed was the real East Kilbride Expressway, but how is anyone supposed to know? Well, here is the route of the actual East Kilbride Expressway. I found it in a list of motorway exits in the United Kingdom. In addition, in 2009 it was cited on the Internet as the location of a UFO sighting. How can you be more authoritative than that?

There it is in all its glory

Doesn’t Google ever look at this stuff? Oh yeah. That’s right, it’s totally automated. The way they find out their maps are wrong is when one of the stupid humans using their database tells them they have an error. Given the tutoring provided the UK population by the Ordnance Survey, I’ll be that Google is going to get a lot of feedback. However, I suggest that you send you contributions to OpenStreetMap, an organization that seems to care about getting it right in the first place.

But how could Google get so messed up – especially when Santa is going to be relying on them?

Oh, did you known that Google is getting ready to toss out the Tele Atlas data it has been using and launch its own database of the UK, so it has been loading its own data in anticipation of the event? You just have to love it when conflation runs wild. However, Google appears to have reached some new lows this time, apparently mixing data from fr.wikepedia.org and other sources to augment its knowledge base of the UK.

For example, how about the M25 (the orbital around London) being renamed “Autoroute britannique M25”?

Autoroute britannique? Hmm

And here it is again next to Heathrow

Maybe near Orly but not Heathrow

And how about this treatment of the Victoria and Albert Museum using a Spanish source?

Si

I don’t know about you, but I can hardly wait for the actual changeover. Maybe Google should try using OSM data? However, it seems very curious to me that the Google Vans had to be using some reference when the drove the roads above and I am sure that the reference did not name each of these roads “East Kilbride Expressway”. Sometimes you can use so much technology that you fail to observe some of the basic rules of map compilation. This is sort of like the snowman that got two lumps of coal for Christmas, put them on his face and said “I can see”. Yes, Google, you can see, but can you understand?

You can blame the above examples that I use to poke fun at Google on a source who requested his name remain anonymous. The source provided the hints, I provided the details. Thanks again for writing.

Now onto some other minor things that might be of interest

My colleagues and I just finished up our assignment with the Geography Division of the U.S. Census Bureau, which we had been working on most of the year. It was great to put the references to bed on the last deliverable. This was the first time, at least as a consultant, that I had worked with a group and I learned a lot from my colleagues, though, once again, we will be splitting up and going our separate ways. Of course, now that I do not have deadlines looming over me, I will turn to writing several pieces that I have been thinking about for Exploring Local, but just did not have time to write.
Frederick Ramm sent me a copy of his new book on OpenStreetMap (OpenStreetMap – Using and Enhancing the Free Map of the World – Ramm, Topf and Chilton 2010) and I plan on reading it over the holidays and reporting on it sometime soon. It looks quite interesting! Two of our five reports for the Census (in support of the 2020 Census) had major sections on crowdsourcing and OSM was a focus of these sections, so I am anxious to continue my exploration of OpenStreetMap.

NAVTEQ News

I had hoped to be able to write a blog on the new NAVTEQ vans jointly fielded with Bing, but the opportunity did not turn out as I had planned. You may remember that I had asked NAVTEQ, though a plea in one of my blogs, for a ride in one of their new vans. You know, the ones equipped with Dilithium Crystals.

Much to my surprise in October I received a call from the PR group NAVTEQ uses telling me that NAVTEQ had though it over and wanted to offer me a ride when a van was in my area. We left it that there was supposed to be a van at a show in Irvine, CA in November and that they would contact me at that time to arrange a ride.

Well, I sent the Public Relations group an email at the end of November and asked what had happened to the ride. I did not receive their first answer, although they later said that they did send a response but that I had not received it because they did not send it to my email address (always helpful). When I contacted them again, they realized that they had sent the email to the wrong address (guess they did not use the reply option) and told me that “Unfortunately NAVTEQ is not currently offering Ride and Drives in their new vehicles.” Guess that was why they sent the message to the wrong email address.

I thought it was curious that NAVTEQ’s mouthpiece said “yes” and then “no” at a later time. Following my first email I wrote a more pointed letter because I thought I was being ignored. I received this response. “My colleagues NAVTEQ would love to have you ride in one of the cars when one is available. Right now, however, there is not a car available. When we do have one in your area, and we’re able to schedule something, I promise to contact you. Let me be clear about the fact that this is not a “No” indefinitely at all. This is just a delay due to scheduling. “

I loved the part “…this is not a “No” indefinitely at all.” How do you parse that? It is like “…this is just a “No” for the remainder of your lifetime”? About the only thing missing from the note is the line “…when Hell freezes over.” Well, thank you for your consideration NAVTEQ, hope that works out well for you. Let this be a lesson that even if you heard what they said the first time, it really wasn’t what they decided in retrospect that they had said to you to begin with.

Merry Christmas to all and to all a good night.

Mike

Bookmark and Share

Posted in Authority and mapping, Google maps, Mapping, Navteq, OSM, TeleAtlas, map compilation, openstreetmap, routing and navigation

8 Responses

  1. Thom

    Please tell me you’re being facetious when you say…. “I suggest that you send you contributions to OpenStreetMap, an organization that seems to care about getting it right in the first place.”?

    Hi, Thom – thanks for taking the time to comment, even though my Christmas Cheer is draining away at an alarming rate due to your vicious attack. I will be sending you coal for your Christmas present – (See, that’s an example of me being facetious, as I actually think your comment was appropriate.).

    I realize that OSM has many flaws, but the reason that people contribute is, I think, because they want to create a free, accurate map of the world. It appears that the OSM coverage and quality in Germany is pretty good, and the OSM coverage and quality in the UK appears to be improving The US coverage lags far behind. However, I don’t think that the current state or quality of the coverage can be taken as an indication that those who contribute to OSM don’t care about getting it right in the first place. The nature of crowdsourcing is that people want to help out.

    Yes, I do agree that OSM will forever battle the problem of being represented by the crowd. While over time the errors should diminish due to the self-healing nature of the process, at any one moment OSM can be as bad as………..well, as bad as the blunders Google made on the East Kilbride Expressway in … well, all over Scotland.

    Thanks again for your comment.

    Mike

  2. Thom

    When YouTube first began, it was the first significant medium that allowed anyone to upload videos they created. When word got out that they were actually paying for “quality” content, YouTube became an Internet sensation. Apparently people thought they were going to be the next Spielberg. However the consequences of allowing people to upload “centerline” data to OSM is doggedly more dangerous than not getting a thumbs-up on YouTube.

    I don’t think most people are aware that OSM collects more than just streets. They’re essentially a collection of motivated individuals that lack the geo-technical wherewithal who are compiling foundation data. I’m sorry Mike but I demand much more from my data.

    Putting a gun in the hands of an untrained soldier is going to get people killed.

    Don’t get me wrong, I am all for innovation. Having location-based services as the number one subject on the tech blogs makes me proud to be in the industry. The use of crowd sourcing and collaborative mapping in Haiti has quickly become the benchmark for that application. However when US agencies like USGS and NGA started moving towards the use of OSM and commodity data as a source for their products, I feel that this trend has gone too far.

    The integrity of our geospatial data is becoming diluted with every innovation and application in spatial technologies.

    Thom

    Hi, Thom:

    Thanks for the comment. You make several good points, but your conclusion that “The integrity of our geospatial data is becoming diluted with every innovation and application in spatial technologies.” is reprising Chicken Little’s “…the sky is falling.” While I can understand your reluctance to use OSM data, surely you do not think that every innovation in spatial technologies is diluting the value of all spatial data. And speaking of spatial data, do you actually think that the spatial data of the USGS, NGA, or anyone else is perfect?

    I think the interesting issues is “Can OSM close the gap between where it is today and where it needs to be in order to be considered the equivalent of data sources that meet your requirements for data quality?” OSM and other crowdsourced systems are considered self-healing over time in terms of data accuracy. We are in the period where this belief is being tested and time will tell if improvement is possible without alteration of the OSM model.

    However, you may remember that for almost all of 2010 the major navigation mapping companies were publishing incorrect data for the now-famous realignment of I-195 in Providence, Rhode Island, well after OSM had portrayed the road geometry correctly. If you were to look at USGS data and NGA data, I suspect that you would find that this interchange remains incorrectly represented on their products and databases.

    I believe that the use of crowdsourced data to serve as an indication of change detection is an excellent method for targeting map compilation research. Tele Atlas does this using TomTom Map Share’s active and passive community data. NAVTEQ collects crowdsourced data for the same purposes with its Map Reporter and through the use of probe data capturing the paths of vehicle fleets owned by companies using its data.

    While you may be willing to impugn those whose collect OSM data, I am not of the same opinion. My background in mapping informs me that the majority of people who collect and compile map data for major mapping companies and government agencies had little to no experience with the collection of spatial data before they came to these organizations. In essence, when they started they may have had experience using spatial data, but it was the rare person who had ever actually collected data in the field or had experience with analyzing data sources to determine which was authoritative and accurate.

    What makes the “big boys” different than OSM is that a corporation or government agency is willing to train its neophyte data compilers on how to discover and capture the quality of data required to meet the specifications of their databases. The question, then, becomes, “Can OSM or other crowdsourced efforts ever hope to equal the spatial database quality of those organizations that are willing to pay for excellence across the geographical areas they cover?”

    The crowdsourced model, of course, does not involve pay for performance. On the other hand, the crowdsourced model does also not have to pay for the management layer who manages data collection, as happens in industry in government. As the world’s economies decline, it becomes less and less likely that these “professional” organizations can continue to spend the amount of money they have spent in the past to maintain their data coverage and data quality. If they do not turn to innovative uses of crowdsourcing, I fear that their data will soon become out of date and of poor quality.

    Unlike you, I applaud NGA, USGS and other for considering crowdsourcing and how it could be leveraged to the advantage of these agencies. USGS, for example, has experimented with the OSM technology stack and with cooperative editing between national and state agencies. I think it unlikely that they would ever be able to adopt crowdsourcing as their sole compilation technique, but believe that adaptions of it may be of value to all mapping organizations in the future.

    Next, I was surprised by your belief that most people are not aware that OSM collects more than just streets. I suspect that most people who read this blog know the details of OSM data and how it is collected, I certainly do.

    Finally, I do not expect OSM will replace NAVTEQ, Tele Atlas or the spatial data provided by numerous government agencies around the world. Quality mapping generally requires a significant investment in data compilation and quality control. It is the financial willingness to support the infrastructure required to meet specific levels of spatial database quality that eventually determines actual quality of any spatial database. With Google giving away map data for free, the real question is whether the levels of map accuracy we have come to accept will continue at this level in the future – or whether “just good enough” will become the new hallmark. If just good enough becomes the standard, will OSM meet the bar? Time will tell.

    Thanks again for your comments, as I enjoyed hearing from you and writing a response to your thought-provoking assertions.

    Mike

  3. Thom

    Your response was very enlightening and provoked me to respond once again. My biggest concern is not necessarily the actual street data; it is every additional feature on the ground that OSM and other organizations are compiling. I see OSM “cartographers” hit the trails with GPS receivers all the time. I stop and ask them what datum are they compiling against, or what collection interval have they set. They look at me as if I am speaking Greek. No wonder why I see 10-mile trails on OSM with only 20 vertices.

    Now you don’t need a geography degree to trace a line on a computer screen from 3year old IKONOS. However a true cartographer boasts comprehensive knowledge of just about every scientific discipline to insure a quality product. Of course they were hired with little or no experience, but years of training has put my faith in their databases.

    I would never say that government mapping is perfect. There is no such thing as a flawless map. You and I both know that the very second a map is produced it is out of date. However the government has thousands of pages of standards it is ruled by, along with a vigorous campaign of quality reviews and training. This structure will produce a product that will be made to the highest standards using all practical resources available. It is not perfect, but it is far better than what collaborative mapping can produce.

    I don’t think OSM is ready for prime time. Not until they adopt standards and acceptable quality measures. However as you so astutely pointed out, whose quality measures? Is our society so used to the current quality standards that people won’t or don’t care? Is sacrificing quality worth the money we will save? Time will certainly tell. I’m curious to know what our troops think of that? Perhaps I am just barking up the wrong tree. I just hope people realize that they get what they pay for.

    Thanks for the responses Mike. You seem to be one of the few people in our industry who get it.

    Thanks for your comments, Thom and my best wishes for a Happy Holiday Season.

    Mike

  4. Dave F.

    “I just hope people realize that they get what they pay for.”

    With OSM you get out what you put in. If you don’t like what see, you can improve it. You don’t have to wait years for so called well trained experts to get their act together. Ordnance Survey has field data that is around fifty years old, & it wasn’t very accurate back then.

    “The integrity of our geospatial data is becoming diluted with every innovation and application in spatial technologies.”

    Pure protectionist babble, Thom. Scared of loosing your job?

    “I don’t think most people are aware that OSM collects more than just streets.”

    You say that as if it’s a bad thing!

    “They’re essentially a collection of motivated individuals that lack the geo-technical wherewithal who are compiling foundation data.”

    I’m not sure I need geo-technical wherewithal to map my local Burger King & a good, strong foundation is exactly what OSM needs.

    OSM, is more up to date than any map I’ve seen, especially in urban areas. New roads & developments are added as they’re completed. Show me a map that has the ability to show the opening times of pharmacies?

    Hi, Dave:

    Thanks for your comments. Since this appears to be a disagreement with the comment by Thom, I will leave it to the two of you.

    Happy Holidays,

    Mike

  5. Ackthpt

    Where is your free world mapping project? Pick pick pick. It’s a free service dude, you aren’t paying for it, and considering the scale, it’s pretty f-ing accurate. Quit whining.

    Thanks for your comment Ackhpt (I did not make that word up; it’s the name on the comment).

    I’m not sure whether you are referring to my comments about Google Maps (which, as far as I can figure is free or perhaps less than free if you license the Google Maps API and run Google ads on the maps), or my comments about OSM (which, also is free and requires a license to use).

    Next, I’ll whine when I want to about map accuracy. According to my records I have complained about the map accuracy of every map database I have ever encountered. If you have to ask why, then you need to spend more time researching map accuracy. I recommend you read my blogs to help improve your understanding of map accuracy and why pressing the map database community to improve it is good for everyone.

    Now your comment “…considering the scale it’s pretty…accurate”. Since, as noted above, I could not determine to which map database you were referring, I am not sure you would find my response interesting or perhaps relevant to your concern.

    Did you know that I once drew the most accurate map of the world on a quark, using a Dill pickle dipped in a can of cooking grease stored in my refrigerator? I labored mightily to ensure the accuracy of my representation and defy anyone to prove that my quark map of the world is not the most accurate map of the world ever made (due in part to the unusual “Peters quarkcator projection” I ginned up for the occasion). Of course, being a kindly person, I released the quark back into the wild, so it may be hard to find. Did you really mean “map scale” or were you referring to the scope of the challenge of mapping the world?

    Finally, if you can’t compose your comments without using terms like “f-ing accurate” don’t bother to write, as I am unable to determine the relative magnitude of these terms and how they meaningfully interact with things like scale and map accuracy.

    But thanks for taking the time to leave a comment.

    Mike

  6. Federico

    Dear Mike,
    I came to this post by following one of those underlined blue words I found on another page, I don’t exactly remember which one it was and I can’t help but leave a letter-like comment on this so-called blog.
    Don’t be mad at me if I mock you a little, it’s just because I’ve heard that comments don’t have to be answered one by one by the post’s author – but don’t quote me on that, I’ve only seen that behaviour all over Internet.
    I found your last reply hilarious and, being the Internet user I am, I just had to leave this comment.

    Best regards,
    Federico

    Thanks for you comment Federico.

    I read your comment and am not sure I could discern the point of your note. On the other hand, perhaps you did not have a point and were just writing me a fan letter. The reason that I am responding is that yours is the first fan letter I have ever received.

    Mike – so called author of this so-called blog.

  7. Anonymole

    East Kilbride Expressway is a thing not a location. If you searched the web proper for that term you would no doubt locate it. But on a map, it’s like searching for I-95 (Eastern seaboard), or the Trans Alaskan Highway, those are not places, those are proper nouns. I’m not surprised that GMaps failed to find your “location”.

    Now, however, I do think that such a thing SHOULD be locatable on a map. Other “things” are searchable on a map, The WhiteHouse, the Vatican, The Eiffel tower… But when it comes to the things that are half location and half thing, I’ll have to say that we’ll be waiting for the next semantic layer applied to GMaps to allow us to find such quasi-location items.

    Thanks, Anonymole.

    It appears that the first figure in my blog about the East Kilbride Expressway did not the intended message across and I am unsure what more I can say to remedy your incorrect observation, but I’ll try.

    If you look at the first figure in the blog, you will notice that the search engine for Google Maps not only recognized the term I entered (“east kilbride expressway”), but also displayed a disambiguation table showing ten instances of “E. Kilbride Expressway” that were discovered in the Google map base of the United Kingdom. Further, Google Maps was able to display (map) each and every instance of East Kilbride Expressway that it discovered.

    As to your assertion that East Kilbride expressway is not a location, I disagree. For example, when someone tells you to travel east until you encounter Route 57, the road serves as a location. Further, East Kilbride Expressway is an attribute of a sequence of vectors in Google’s map base for a location near Glasgow, Scotland. As such, it is part of the database that creates Google Maps. Whether Google chooses to expose these data everywhere is their business, but in the case at hand, the East Kilbride Expressway was found by the search engine in Google Maps. The problem highlighted in my blog was that in the majority of the instances the East Kilbride Expressway discovered by Google did not exist in the road network of Scotland (not the imaginary one in Google Maps, but the real one, on the ground, in Scotland).

    Thanks for your comment,

    Mike

  8. Victor

    Now that Google local results (Google Places) are appearing on Google.com search results, getting it wrong on local has a much broader impact. Case in point, search for “best hotels in new york” and four of the first seven results are Best Westerns! The “normal” search results would never return a BW on the first page.