Open Maps?

Chilltowner asks: "I'm trying to get local (US) maps together for a community project. I want to able to modify and annotate the maps and provide them free to the public, creating a derivative open work. They also need to be accurate down to the street level and no more than 10 years out of date. I've been searching around for maps available in the public domain or under open licenses, like the Creative Commons licenses allowing derivative works. I've looked at the National Atlas, but the maps, though interesting, aren't detailed enough with street information. The topographical and aerial image maps available through that site are from Terraserver, which are copyrighted to Microsoft. Plus, I really just need simple vector road maps, not USGS rasters. I tried looking at the Census Bureau's TIGER line data, but I can't make heads or tails of it. Are there maps available through other agencies (national or international)? Are there Free/Open-Source Software projects that are making use of public data to build street-level maps for free (as in speech) use?"

Comments

Maybe it's time

[lawpoop]

I've seen several projects where people use their PDA/GPS to map their daily route. Maybe it's time someone organized a collective mapping project, for release cunder the creative commons license.

Re:Maybe it's time

[snillfisk]

We've already done a bit of work on this, although not in any open projects yet. We do however have a small testbed for integrating and designing datasets in a SVG-client before going through a peer review process by others that have data in the same area.

Our white paper on the architecture may be seen at SVG Open from the 2003 proceedings. A small amount of information (including a live demo -- it does however require that someone accepts the data submitted, so don't expect to see anything in the gateway at once) is available here.

The idea is however a good one and is absolutely doable. The problem is that you'll probably get rather large datasets for the most used paths, while the roads noone travels frequently will remain unmapped (and those are really the roads that you _do_ need a map for)..

I'd suggest really old maps

[Anonymous Coward]

They're out of copyright by now, and unless you're in one of those fancy "new" cities like Phoenix, they'll probably be pretty close to how things are now, or at least easy enough to add to.

Re:I'd suggest really old maps

[line.at.infinity]

I've seen some old aerial photographs of the city I live in (and I wouldn't be surprised if those photos were as old as webster 1913) but the major roads, rivers, and landmarks still remain in the same place. While we're at it, we could make a digitized map of the past (maybe some historians with grant access would be interested) and edit on top of these maps. The one problem would be the lack of GPS information from old maps, but that could be solved by extracting coordinates from the maps and photos, with probably reasonable accuracy.

The key to a copy-free solution would be maintenance. Just copy how the major map companies update their data. And the public would do a better job of it since as a whole the general public has more itches they want to scratch than the few paid workers who update maps. (e.g. "that road doesn't exist!!") As mentioned on /. way back when, one could drive around with a GPS coordinates recording device.

Re:I'd suggest really old maps

[cdrguru]

You can forget about accuracy with digitizing from aerial photographs. NavTech tried this for a long time and there are just too many corrections that needed to be applied to get things right. The photographs they were working from were paid for to be done "right" and they still didn't have the needed accuracy.

They now collect all geometry with differential GPS in car - driving all the roads.

Trust me, you aren't going to get anything usable from aerial photographs that were not created with this application in mind.

Making maps is not an esoteric science

[Anonymous Coward]

People have been making maps for hundreds of years, maybe thousands of years.

Hell a Japanese guy with no formal mathematical training was able to figure out how to make very accurate maps (especially considering the poor accuracy of the maps of Europe) using no more than 300 men, several teams of horses, and large sextants and compasses.

Why don't you start up a mapping project on your own and put a subproject idea under the main banner encouraging people to implement whatever harebrained scheme you are talking about. The community will enjoy your work and you will gain notoriety as the guy who opened maps to the world.

PostGIS

[SEWilco]

First, look at PostGIS. It is a geographic extension to PostgreSQL. That gives you a single place to store your data.

Then look for "TIGER PostGIS" to find tools which support both formats, and find something to read TIGER into PostGIS. Then look at editing and display tools to find one which supports PostGIS.

Maps and accessories baby...

[chamilto0516]

I was suprised when I bought my GPS unit. The maps (or unlocking the maps that we shipped with it) were almost as expesive as the unit itself. I have a Garmin eTrex Venture and between the Garmin US and DeLorme TOPO USA, I have paid more for this data than the hardware.

The maps are where the GPS device companies make their profit. That and accessories ($35USD for an AC car adapter!).

If I were to ever start my own Open/Free project, it would most likely be a call to all us GPS hobbyist out there to create our own Open/Free maps and GPS coordinates of useful landmarks.

Excellent Ask Slashdot question...

Re:Maps and accessories baby...

[cosmol]

Then as hundreds of people move about, all recording it, then you'll build a map.
That's sort of like the strategy ant colonies use to establish paths. It's interesting, but such a brute force method would duplicate much effort, and miss many routes. The data is out there, we just need to convert it into a usable format.

Try FEMA

[Anonymous Coward]

You could check the FEMA site. I'm not sure what license they use and the maps might just be from USGS.
The maps are intended for flood plain information and I recall some difficulty in finding my street back when I needed to use their system. You can't just enter an address and get a map with a red X. You need to figure out what part of the grid your in and select your panel.

Maporama

[Alien54]

I actually prefer using Maporama, which allows you to generate maps up to 999x999 pixels.

You can choose a number of color styles, and you can save the generated map as a gif file, which is can then edit with common software. Very configurable, and an account is not needed.

They also provide street numbers when you are zoomed in close enough.

Overall, worth a bookmark.

Property Appraisal Maps

[webber1]

Why not check with the property appraisers in the area for which you seek a map. Most are to street level and are pretty up to date since the taxation depends on their accuracy?

Re:*scratches head*

[ericspinder]

Yes, you have "missed something" the National Atlas (a service of the USGS) is clearly public domain material. As is (I think) all published government sources. In fact, most of what the map makers do is based on government surveys and publications. The companies do some fact checking(some better than others), add "features", and consolidate information, but the 'base' work is often public domain.

Re:*scratches head*

[jamestedrick]

Map data produced by the US government is generally available royalty-free for reuse; once Microsoft/Delorme/RandMcnally modify/refine it, they can then apply their copyright to the derived product. In the UK, national mapping data is kept under crown copyright.

GPS calibration

[Doc Ruby]

Where can I find GPS-calibrated maps? Or just GPS calibrations for these other standard map formats? And while we're at it, how about a GPS points exchange, for swapping points, paths and tracks?

Re:HOLY CRAP

[timothv]

If you thought that was cool, wait till you see map24.com

MARPLOT

[Big Sean O]

The US EPA and NOAA have a free (as in beer) computer program called MARPLOT.

It was initially meant to be used by emergency responders as part of the "Computer Aided-Management of Emergency Operations" or CAMEO. It was so popular that the US Census Bureau made it part of the Landview software program.

It's not as nice as a professional program, but there are lots of basic features and the price is right.

County records

[LuxFX]

I live in Johnson County, Kansas, and they have a very impressive mapping system available online. (it is most functional in IE, but Mozilla etc. will let you do the basics)

You can find individual properties (complete with tax appraisal information, square footage, room count, etc. -- did you know this is all public information?), property lines, estimates acreage, building outlines, etc. You can map water pipes, power lines, fire hydrants -- even many trees are included. Fire stations, parks, museums, streams, neighborhoods (plats), cities, etc. It's all there.

Very impressive! Check and see if your county does the same! I can't tell you how valuable this tool was as we were shopping for a house (we closed yesterday!)

Re:Try the County Assessor yet?

[winwar]

Do you really need the accuracy and precision you are talking about? For 911 dispatch, who cares which property the tree is in. That level of precision and accuracy would be nice but is it needed or would the money be better spent elsewhere. Hell, USGS 7.5minute quads combined with Tiger, orthophoto, parcel data, and other info is probably good enough. I mean, how accurate is that water main map anyway.... After all, the responders have brains and in theory have at least a passing knowledge of the area...
Finally, in theory property lines used to create parcel maps were surveyed and you could just digitize them. Then georeference them. It should be good enough. If it isn't, then I doubt any amount of GPS is going to help at anything resembling a reasonable cost (if the maps are good enougth for people to find hydrants, water mains, property-then that can be transferred to the computer). Granted, it isn't easy. It's boring. It's tedious. It's time consuming. And the surveys and maps aren't always correct. But don't assume guys pounding pavement with a gps will do any better for anything resembling the same cost. Survey grade gps does not equal absolute gps accuracy. More correctly, you would need guys driving a van with differential gps to do better for roads at least(Navtech comes to mind)...

Maps are not copyrighted

[tiltowait]

so go nuts with whatever you can get your hands on. At least that's what the law was the last time I checked: you can't copyright a fact (or a made up fact for that matter), although some people are trying to change this.

I had a the pleasure of once working for a map company, for example, that at a time (before I worked there of course) traced a competitor's maps when drafting their products. An ensuing lawsuit, during which the judge actually acknowledged this practice, resulted in a verdict in favor of allowing such infringements.

Re:Maps are not copyrighted

[Frogg]

Maps likely have different copyrights depending on the map in question and what coutry's copyright laws are applicable to the map-bearer.

Here in England we have "Ordnance Survey" maps which are very high quality -- and are all covered by Crown Copyright (the government/queen's copyright -- I dunno how it's any different to any other copyright though).

Also, if you take a map to a copy shop here in the UK -- let's say, because you need to give your dad (who lives some way away) directions how to find your street -- many of them will tell you that they will not copy it for you (bah, just like when someone in your band needs another copy of some sheet music).

I have a feeling that the "Times' Atlas of the World" is also likely to a copyrighted work..!

Re:USGS

[Technician]

Don't overlook commercial software. Much of it uses common data, but provides a nice user interface and features that may make the price reasonable. The price for map packages are falling rapidly as the early adoptors that will pay anything have already paid. Now the rush is on for market share. Take advantage of the corner gas station price wars. There is a war between Microsoft (no suprise) and Rand McNally for vector street maps. Microsoft is trying to keep the percieved value high by keeping the list price high but it trying to get market share with a heavy rebate. Street level vector maps are in the under $20 range for ones that work well with your GPS and PDA. Delorme got skunked on the unexpeded fight. They tried selling a map, then having the GPS and PDA stuff as seprate packages. Ooops sorry Delorme, you missed the boat.

On the TOPO side the war is between Delorme, and Natiional Geographic. Delorme sells a vector based topo map which is nice. National Geographic sells Rastor Maps of the USGS maps. The state series is a little spendy, but detailed. It's a clear winner if it is out for your state if you do off road and back road stuff. However having a picture of the streets is it's weak spot if you are trying to use it for street navigation. In a nutshell, if you use the TOPO maps by National Geographic, you will want another map for street use.

Just out from National Geographic is an answer the vector street problem. The Back Roads Explorer map combined the Raster Maps with a real time overlay of the street vector maps. The 17 CD set can be bought for under $40. It contains the entire USA TOPO and vector street maps. It's detail in the TOPO isn't as good as the state series, but you can load the state series into the package for the best of both.

Now the issue of marking them up and such. National Geographic permits printing of maps provided the copyright remains on the map (in the boarder, not imbeded like Mapquest). I'm not sure of the quanity or uses of the maps so more research will need done especialy for commercial use, but home printing and sending maps to your friends in your hunting party all marked up to the hunting camp is a permitted use. The tools to mark up the map are included and work well. Added to the ability to export to my PDA or handheld computer and connect to my GPS and upload, and download waypoints, routes and active track are all features that make this commercial map a winner.

Disclaimer, I don't work for National Geographic. I like to Geocache and finding a road in close to the cache is half the fun.

Mapserver

[ThesQuid]

There's a great open source project that you can use the tiger data with - Mapserver
Lots of tutorials there to, but it's a bit of a learning curve. Try it! I knew nothing about mapping, and in two months I had built a web application that could zoom down to the street level with selectable layers for all sorts of data.

Re:USGS

[spatialguru]

That's a good overview of the options and issues, thanks.

The point that seems to be missed in the threads is that (as any fellow geocacher knows) street maps and topos just don't cut it for many applications.

Existing street mapping, etc. has some mass appeal and real utility, to be sure. But at many local levels, small private groups has whacks of spatial data that is never shared - and partly because no one would care anyway.

For example, forestry, environmental, regional governments and municipalities all have map reporting requirements (i.e. to gov't, public) - leveraging that data into the public domain provides excellent and more meaningful local information.

In short, folks need more than a good MapQuest. Show my the location of my house on a map and I'm slightly interested. Show me the path through the woods by the neighbour's house and I'll be impressed. Show me a localized plant hardiness map and I'll be excited.

Consider a geocache located in a landfill - Delorme ain't gonna warn ya! :) -Tyler

Re:Since you're in the US..

[PantsWearer]

Since OpenGIS was brought up, I figured I'd chime in. I've been working with various OpenGIS standards for the last couple of years, so I'm familiar with them and also have some opinions about them.

WMS (Web Mapping Service) is a pretty good standard. It's simple and easily interoperable since you basically can type in a URL in your browser and come back with a map. Of course, since you get a static image back, you have to make another round trip to the map server to do any manipulation at all. In other words, it's really easy to use (and to implement), but it's use is both hard on your network and on the mapserver.

WFS (Web Feature Service) is a vector format for maps. In theory, this is really wonderful. You request the map area you want to see, a ton of vector data is downloaded to your viewer (SVG or other vector drawing tool) and you can zoom in and out as long as stay within the area you first requested and don't want to see anything new. This is a little bit harder on your processor, but with today's machines this isn't much of a problem. And it eliminates the round trip problem which WMS has.

It has it's problems though. A minor problem is that it doesn't support raster data (such as aerial or satellite imagery). This means, if you want to overlay aerial photos you'll have to hit a WMS for the photo itself and then put your vector data on top of it.

The major problem is the GML format itself. This is how a WFS sends data to you. It's a particularly verbose XML format that, as of the current standard for WFS, is uncompressed when shipped. Geographic vector data is data dense. As a simple example, say you want to give the lat/lon for every house in your neighborhood. You'd probably end up with a list of about 60 doubles (assuming 30 houses). Now print all that out in a text format and surround them with about four or five levels of XML tags. Not that you could easily write this up yourself: the GML specification is an incredibly dense, 300+ page document.

Basically, GML is an impractical solution because of the opaqueness of the standard and the sheer size of the data it produces. The standard's opaqueness isn't so bad, since more or less GML is only machine readable, but the size is a major roadblock.

I'm not saying that there are better standards out there. For example. ESRI's Shape file (an industry accepted "standard" which ESRI, who is the Microsoft of the GIS world, created) is one of the nastinest binary formats I've ever seen (why would you have a binary format that contains data in both little endian and big endian at the same time?), but at least it's size friendly. For example, a shape file of the state boundaries of the US might be about a megabyte (this would probably be only 1:1000000 scale or slightly better, so once you get down to a county level, the state boundaries would be really choppy). The equivalent GML file would be about an order of magnitude larger, so if a person wanted to view something of the sort in their browser without having to continuously download more detailed data (the WMS problem again), they would need to download about a gigabyte of data just to see the state borders and allow for some fairly high level panning and zooming.