[Talk-us] Data sources (Was: Proposal: Sunset ref=* on ways in, favor of relations)
kennykb at acm.org
Wed Nov 11 15:23:51 UTC 2015
On 11/11/2015 06:53 AM, Paul Johnson wrote:
> I have Phil's code adapted to my own use, and going at
> (that particular spot shows Interstate, US Highway, NY, NJ, PA and
> Wow, that's a stunningly pretty American-style map there, and one I
> wouldn't mind seeing on a potential map page on openstreetmap.us
Well, thank you! I'm glad you like it. This started from Lars
Ahlzen's TopOSM project from a while back, but I've taken it off
in a somewhat different direction.
It's something of a personal project that's got entirely out of hand.
I realized that if I am ever to have useful maps for hiking less
popular areas that properly support panning and zooming on a smartphone
GPS app, I have to make them myself. The publicly available USGS maps
are decades out of date. The USGS topographic mapping project ended
in the first Bush administration, and mapping had been done on a 40
year cycle prior to that, so the maps are for the most part between 25
and 65 years old. The newer 'USA Topo' series is nearly worthless
for the purpose (and USGS knows it). And outfits like NatGeo and
DeLorme often don't cover some of the, uhm, obscurer places I go.
I don't know how appropriate the map will be for openstreetmap.us,
because I do depend heavily on quite a fair number of non-OSM data
sources. In particular:
- NED - both hillshading and contour lines derive from here
- NLCD - It's low resolution, but I don't think we're ever going
to have crowd-sourced landcover information that's any
good. And I want to know when I'll be filling my boots
with beaver water or ripping my gear on spruce.
- NHD - I've heard many complaints here about NHD, but in the
Northeast, its data quality is quite good indeed. This
isn't a project that armchair mappers can complete
easily, and access to land for field checking is often
difficult or impossible. I'll go with what the
government has on this one. (I ignore hydrographic
features from OSM).
- USFWS National Wetlands Inventory, plus NYSDEC Wetlands files,
plus Adirondack Park Agency's commissioned wetland studies.
Coverage varies here, and the files are intended to serve
different purposes. Since whether a property is legally
a wetland or not has vast impacts on its value,
this information is always contentious. I'll accept
information of varying quality here. The key point for
me is "how likely am I to be wading?"
- Several states' databases of hiking trails. The legal status
of New York's is murky (I gather because the State doesn't
own the data, either - it was provided by a contractor. It's
nonetheless available to the general public at the NYSGIS
web site, and I'm comfortable with the risk for my personal use)
The other States that I've included are much more open with
Particularly near the corner of NY, CT, MA, this gives the
trails something of a 'cubist' appearance, since there can be
as many as four data sources supplying them. I actually find
this valuable. Where the sources do not concur, a trail is
likely to be more difficult to follow, and I'm going to be
paying considerable more attention to navigation.
And there are probably a few I've neglected to mention.
Is there policy on rendering based on data from outside sources?
It also, at the moment, needs a patched Mapnik to render. Phil!
Gold's stuff doesn't work on a read-only database connection, and
Mapnik sensibly requests one.
There may also still be a dependency on an auxiliary table in
the database that Lars used for merging route and way information for
highway shields. I think I've expunged it, but I'm not sure.
It's all slightly "chewing gum and baling wire" - what software
isn't? - but if there's interest, I'd be happy to share what I
One last warning. It's a hiking map. It looks like hell at high
zoom levels in the cities. Sorry. That's not its intended purpose.
More information about the Talk-us