Open US GPS Data?

tobiasly writes "I read an article today about a map error on the popular Garmin GPS devices which often leads to truckers in a particular town becoming trapped. From my own experience, every electronic map I've ever seen (Google, Mapquest, my Mio GPS) has the layout of my neighborhood completely and frustratingly wrong. A quick search turned up only one open-source mapping project, but it's for New Zealand only. Why are there no comparable projects in the U.S. or elsewhere? Obviously such a project would need a good peer-review/moderation/trust system but I'd gladly put in the time necessary to drive around town with my GPS in "tracking" mode, then upload, tag, and verify my local data. Has anyone with more technical knowledge in maps and auto-routing looked more into this? Are there technical limitations to such a project? Should the government subsidize a project to create open, free, up-to-date electronic maps? Surely there is a public benefit available from such a project."

That's a pretty big job

[Exp315]

The only two suppliers of nav map data in North America are Navteq and TeleAtlas. They have both invested huge amounts of money in creating their maps, including driving around cities doing street-by-street mapping with vans, although most of their data came originally from official public street maps. Both companies have been the target of multi-billion dollar take over offers in the last year. In addition to capturing the map data, tagging (street names, one-way, turn restrictions, road type etc.) and validation (making sure streets link up correctly in the database) are also huge jobs. I wouldn't want to say that an open-source effort is not possible, but we shouldn't underestimate the magnitude of the job. It involves a lot more effort than just driving around a few streets in your neighborhood.

Government Maps - of course

[whm]

The government already creates these maps (TIGER [wikipedia.org]), which are in the public domain. But I'll admit, it's a little fun to pretend that Google/MapQuest/Yahoo and whoever else are driving around all of the Western world with GPSs attached to their cars :)

~whm

Re:Government involvement

[fistfullast33l]

You think adding the Government would help improve mapping products? I'll keep my tax dollars, thanks.

I would point out that Government funding is the reason that you are able to A) connect to thousands of computers/websites across the globe right now, and B) the reason that you even have a "computer" sitting on your desk. Ironically, this funding is also the reason that satillites in space can provide us with overhead images that you see in Google Maps and the like as well.

Re:That's a pretty big job

[Jim_Maryland]

Building a community based dataset may have some benefits but it also has many problems. The benefit is that it will be an open source of data and anyone can provide updates to the data. The downsides are: - enforcement of attribute: either people must be forced to enter certain attributes to ensure consistency in the data (which will cause some to not participate in collection). Without this, the data can not be used for more complex usage (geocoding being a primary problem for typical web usage) - accuracy of entered information. Misspelling names, wrong type (Road, Street, Highway, Court, etc...), address range (do you consider it as a "hundred block" or just list the actual house range, i.e. 401 - 438 vs. 401-499), do you consider addressing the sides of the street (left range 401-499 and right 402-498, or if a "T" exist, do you break the one side at the "50" mark or whatever the physical addresses are?) - positional accuracy. Not all collection devices will be equally calibrated. - collection focus. Will data in more rural areas be collected as often as city blocks? These are just some of the issues that come to mind when I think of building a road dataset. I've worked with development of enhanced 911 datasets for counties in the past. These have also used for other departments like the school board (school bus routing), public works (trash collection, snow removal, etc..), emergency services (response districts for police, ems, fire), and other uses. What would really jump start a project like OSM would be the donation of a dataset from a company like NavTeq or some open source group that can negotiate with the local government agencies to provide open access to the centerline data the agencies already have.

Re:Government Maps - of course

[Anonymous Coward]

*cough*still useless to a majority of people because its only in 29 cities throughout the entire US*cough*

Re:open street map?

[Anonymous Coward]

The same could be said about wikipedia when it was OSM's age.

The obvious solution...

[MadMidnightBomber]

Come and live in New Zealand, eh? [newzealand.com]

Re:open street map?

[pipatron]

So fix it. You're obviously a geek since you read slashdot. You obviously have a lot of spare time, since you read slashdot. You also know about the errors and how it's supposed to be. Give an hour of your time to the project. The more complete it is, the more people will fix the details.

The dancing dog observation

[bfwebster]

I'm amused at the thought of trying to create an open-source version of a typical North American commercial GPS street/address database and navigation program. I've used a GPS system in my car for about 3 years now, and while I encounter the occasional error or omission, most of the time I marvel that it works at all, much less as well as it does. As someone who has worked on some very large scale software projects, I have to say that the software quality assurance (SQA) challenges and issues for both the database itself and generating navigation routes from Point A to Point B are enough to give me the heebie jeebies -- particularly given the IT industry's general track record on SQA practices.

Here's a reality check. Pick any one-square-mile area of your community and attempt to create (and keep up to date) a GPS navigation system that will legally, safely, and efficiently navigate you between any two addresses within that square mile, keeping in mind your civil liabilities should your system cause accidents, injuries, or illegal driving maneuvers. Oh, and your navigation system has to fit in a device that's about the same size as a Palm Pilot or an iPod touch and that runs on rechargeable batteries.

Now scale this up by about 3.5 million to cover the United States. ..bruce..

Bwahahaha!

[Alex Belits]

The electronic maps don't show a gate that separates residential and industrial areas. It's only opened for a couple hours on weekdays in the northern New Jersey city.

Mayor Dennis Elwell says residents on Fifth Street started complaining about trucks clogging their street about a year ago as GPS devices increased in popularity. Some drivers have to call police to open the gate because their trucks are too big to turn around.

It looks like they made a gate to shield some gentrified neighborhood from the contact with lower classes, and ended up with a street full of trucks. Solution: open the fucking gate, you stupid yuppies!

Re:Government involvement

[Anonymous Coward]

Actually, your tax dollars did a fabulous job of "getting" you here or are you using some sort of inertial measurement unit for navigation? Because if you aren't then you are using GPS designed and implemented by the Armed Forces which are paid for by the United States Government which I believe is the sink of your tax dollars.

Re:TomTom MapShare

[Anonymous Coward]

Before you know it, hordes of TomTom-toting patrons will crowd your once lovely restaurant. Prices will skyrocket, portions will shrink and before you know it it will obtain a Michelin star and you will have to find somewhere else to eat.

And if not enough people know about it there won't be enough business, it won't be able to pay its rent, and it will close down.

Re:TomTom MapShare

[Anonymous Coward]

Before you know it, hordes of TomTom-toting patrons will crowd your once lovely restaurant. Prices will skyrocket, portions will shrink and before you know it it will obtain a Michelin star and you will have to find somewhere else to eat.

Most restaurants can't let prices skyrocket. Maybe a handful of 1% could. Chain restaurants won't. Chinese takeout isn't likely to (there are reasonable price ceilings before one goes to the next closest/next best joint). It is more likely that an unknown place with good food will go out of business than be so overrun you can't eat there. If the overrun scenario is going to happen, your intervention is unlikely to hasten the matter. On the other hand, one person can make a difference in keeping a small restaurant in business. 3-4 trips per week at $20 each is $3,640. Some word of mouth on your part can increase that ten-fold. That is one person's salary.

The Reality

[CDOS_CDOS run]

As a mapping professional I can insert some truth to this Slashdot orgy of mis-information. There is publicly available data for road centerlines in nearly every state, this data tends to be rather accurate, and thus large. For the use of GPS, you need to use generalized and often closed formats, and thus this accurate data won't work. Most GPS companies buy their data from data wholesalers Navtec, Teleatlas...etc. I have an assoc who works for one of these companies, they collect the publicly available data and reform it to their formats, they also dither(generalize) the data down so it'll be smaller in filesize. Often times the data warehouses will take the accurate vector data and convert it to low res raster data. If TomTom or Garmin wanted accurate data they could get it, but they are willing to do the work or take the time. Tiger data is very generalized, but will be revised very soon to increase the accuracy. Again this was a matter of data size, the accurate data was dithered again. Also bear in mind these GPS devices still aren't overly accurate. but there is freely available road centerline data in every state I can think of, in my state it is nearly survey accurate. As for street names, that is a whole additional ballgame, that is very intensive data collection.

Re:Frustrating

[westlake]

double it for everything I forgot

Here a few examples I could show you in a fifteen minute drive:

seasonal roads

privately maintained farm roads, service roads, gated communities, government reservations and the like. which share nothing in common but distrust of strangers.

long-obscured, missing or unreadable road signs

names too long for the standard-length sign. abbreviations that are more misleading than helpful

names the locals never use themselves

--- the outer ring of development known since 1934 as "Poverty Ridge."
--- the three block stretch on the south end of Third Street renamed for a beloved centenarian who died in 1956

In short, it's not that easy-

[uncwjason]

As with some of the other posters that have dealt with GIS in one form or another, so do I. I am a GIS analyst for a county government, and I can tell you that when any new roads are cut, we are the ones that start the ball rolling. So, you can see that over time, lots of governmental organizations are the ones that initally put these roads out there (aka make available for distribution). If you take my particular organization, multiply that by every county and local government in the country, you can see the conflation that occurs when your major road navigation companies try to stitch these together. I dont know the exact number, but some states have different projections for their data, and there's at least one for every state. States that have some kind of non-equal extents (california and north carolina come to mind)usually have multiple ones. Assembling this data takes time and effort, even if it's just updating what they already have.

Another problem is that you dont want every tom, dick, and harry editing GIS data for the masses. Control is key, and there is an implication that the data has been quality checked and will lead you to wherever you go. If you have grandpa out there, logging some points and uploading them to the public, how do you know that he put the data through differential correction and the lines are topologically correct?

One final thing...my county doesn't try to profit off of it, but there's many, many governments that charge some pretty high fees for somebody to go in and buy their data. You think that they would give that up easily, when they're basically making total profit off of the data and we have to maintain it as part of our job? No way.

So my advice is this: there are ways to convert and upload basic GIS shapefiles into your GPS units if you so wish. Check with the local authority to get your best data. Our E-911 system uses it, shouldn't you?