[Talk-us] Rail USA

[stevea]

>Also good data can be retrieved from time tables, division maps, and 
>track diagrams.

That is true, thank you Nathan Proudfoot.  For the Union Pacific, I 
was disappointed that their web site requires a (UP employee?) login 
and password to gain access to their geographic rail data.  BNSF 
seems a bit better (providing "high level" rail maps at a nationwide 
glance), and other railroads are probably somewhere around "you get 
what you get," but please do take Nathan's good advice and seek data 
directly from a rail entity as a good first strategy for obtaining 
track/lead/line/subdivision names, for example.

>I'd be happy to convert the shapefile to an .osm file for importing. 
>It could be broken into smaller chunks to use the tasking manager. 
>Maybe by county. Someone needs to create a import page on the wiki 
>and announce it on the import mailing list.

Thanks, Clifford.  I want to make it clear that it is very early 
"first blush" that we (OSM) examine these Oak Ridge rail data as a 
POTENTIAL import into OSM.  We are still in the beginning evaluation 
phase of examining the data, and no decision has been reached as to 
whether we will or will not formally import them.  I do cross-post to 
the import-us list as a preliminary "heads-up" that at least some 
folks in OSM are considering this.  But we are a ways away from 
beginning a formal import process until we (I, others, please share 
your feedback) examine the suitability of these data and begin to 
formulate strategies about how they may best be used/imported.  Or if 
indeed they or a subset of them DO get imported.

Conversion from zipped shapefile to .osm is fairly straightforward: 
simply unzip, fiddle the geospatial coordinate system to be "proper" 
(I'm still working on this), open in JOSM with the Shapefile plug-in 
enabled, and there you go.  Saving to .osm is straightforward after 
that, but right about here in the workflow is where it might make 
sense to break up data (say, state-by-state).

As for a tasking manager breaking up the data, yes, perhaps a 
state-by-state approach is a good first whack at this.  49 (Alaska 
data are included, Hawaii, no) is a manageable number of 
"sub-projects" but around 3000 at a county level, whew, not so 
helpful.  But before we do that, I think I'd like (us, me, others...) 
to take a look at those 30 or so tags on each "line" and see if we 
can begin to make sense of them and how they might turn into useful 
OSM data.  I believe that's an important sub-task to start to do 
first.  Some sort of a table lookup from the abbreviations used in 
those tags to something that is more human-readable is going to be 
required, I can determine that already.

BTW, the actual track data might actually be better expressed by 
OSM's existing TIGER import of rail data.  (And I think that is a 
GOOD thing, as would remove the requirement for conflation of the 
two).  The Oak Ridge track data, while "there" seem a bit sketchy in 
comparison to OSM's TIGER rail.  Jury still out, evaluation of Oak 
Ridge data continues.

We could use that wiki to have these conversations, but Charlotte 
(while she did volunteer to begin it) has a full plate now.  Perhaps 
someone else could plant the initial seed for a "USA Rail 
WikiProject"?  If you wish to do so, contact me or Charlotte for a 
"Draft 3" that Charlotte, Alexander and I evolved in the last few 
days.

SteveA
California
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20141231/fc680653/attachment-0001.html>