OpenStreetMap
After a couple of months of frantic uploading and tagging, I have completely documented the Pacific Crest Trail (http://en.wikipedia.org/wiki/Pacific_Crest_Trail) from Mexico to Canada. That's 2650 miles of trail through 25 national forests, 7 national parks and 3 states. Some of the trail was already marked but most of that was wrong, showing only where the national Forestry Service had marked the trail on their paper maps and not where the trail was actually built. My data was all recorded by GPS when I thru-hiked the trail in 2007.
I've made it into one relation (http://www.openstreetmap.org/browse/relation/1225378/) and that seems too big for the server as the little preview map in the corner no longer shows the route, but Lonvia's Hiking Map does (http://osm.lonvia.de/world_hiking.html?zoom=5&lat=40.70331&lon=-98.88906&layers=FFBT) though of this writing it still hasn't fetched the last few hundred miles to the Canadian border. There are relations for each of the sections as defined in the PCT data book.
This was my first OSM project, and since I have no other major data source to play with it may be my last for a while.
Discussion
Comment from wallclimber21 on 16 December 2010 at 02:36
Wow, congratulations! Both on the OSM effort and the thru-hike!
Comment from Dion Dock on 16 December 2010 at 23:04
I wouldn't necessarily believe a single GPS track is more accurate than the original data. Hike (or drive or ride) a loop a few times and look at the resulting data. In my experience, they are close but aerial imagery is better.
I don't know whether a super-relation would solve your problem or not; they seem to be a work in progress.
Why not upload shorter hikes too?
Comment from UberHiker on 16 December 2010 at 23:53
@Dion,
Where the data was imported from the NFS dataset, or something derived from it, the trail was very much an approximation. Instead of the clear switch backs that my GPS recorded (and can be seen in imagery) it might do one large sweeping turn. I did my best to ensure the final data in OSM was true to the trail on the ground, even when neither NFS or GPS had it recorded correctly.
The few shorter hikes I've done have mostly been covered already. I might verify some of New Zealand's Great Walks when I get the chance but right now I feel I can take a break from mapping and enjoy summer.
Comment from cyberhobo on 19 December 2010 at 06:04
Really an awesome contribution! I've been working on the CDT for quite a while, but I don't have complete data. I'd love to know a few details about how you collected yours.
Comment from UberHiker on 19 December 2010 at 06:10
I just carried a GPS logger made by http://www.ohararp.com and some solar panels to keep it powered.
Comment from UberHiker on 19 December 2010 at 06:15
I see your CDT progress, now that's a hike I'd like to get my feet into. How much of the trail do you have? I'm surprised the AT hasn't been completed yet.
Comment from cyberhobo on 19 December 2010 at 14:53
That looks like a good device. I use a qstarz logger (http://www.qstarz.com/news/news-20080515.htm) with a Brunton solar roll (http://www.brunton.com/product.php?id=256), but I still ran out of power a few times and would have liked the SD card storage.
I have CDT data for Wyoming only (and a few bits in New Mexico). It's an interesting task because the CDT is still commonly accepted as a choice in many place (by through hikers, at least) rather than a fact, so the route can be subjective. A good way to do alternates will be important in the long run.