[Talk-us] importing bus stops

Martijn van Exel m at rtijn.org
Tue Apr 17 00:54:26 BST 2012

Hi Paul,

On Tue, Apr 10, 2012 at 10:52 PM, Paul Norman <penorman at mac.com> wrote:

> > From: Martijn van Exel [mailto:mvexel at gmail.com]
> > Subject: [Talk-us] importing bus stops
> >
> > Hi,
> >
> > Anyone here with experience importing bus stops? Any particular
> > considerations?
> >
> > To make it more concrete, I have permission to import all UTA stops.
> > They come in a shapefile similar to the one available for download here:
> >
> > http://gis.utah.gov/sgid-vector-download/utah-sgid-vector-gis-data-
> > layer-download-index?fc=BusStops_UTA
> >
> > (That particular set of files is out of date, though)
> >
> > There's a number of properties that would map to OSM nodes pretty
> > nicely:
> > MAILBOX -> create a separate node amenity= LIGHT -> lit=yes SHELTER ->
> > shelter=yes BENCH -> bench=yes (?)
> Unfortunately, all of these attributes are null.

Yea, for the public file it was. We sorted this out over IRC. I shared the
file I got from UTA which is more descriptive.

> >
> > I was planning to just use what I know which is highway=bus_stop for the
> > bus stops, and railway=tram_stop for the light rail stops.
> What differentiates them in the shapefile? I don't know the area so I don't
> know where there are any light rail stops

That would be the LOCATIONUS field. It differentiates between:

Bus Bay
Bus Stop
Paratransit Stop
Ski Stop
TRAX station (that's out Light Rail)
Frontrunner station (that's our heavy commuter rail)

> >
> > To get back on topic, if anyone wants to help out devise a mapping from
> > UTA stops file to OSM, I'd welcome some help. I've never done a local
> > import before, and I'm not a particularly big fan of imports, so I want
> > to proceed with caution.
> I threw together a draft translation file at
> https://github.com/pnorman/ogr2osm-translations/blob/master/utahbus.py
> Thanks, how would that work? Do I feed that to an existing application?
This is the first time I see a file like that.

> To go farther, a local would need to provide some knowledge.
> 1. Are there unique ref numbers on the stops? Every bus stop here has a
> unique ID on its sign. If so, do these correspond to LOCATIONID?

I'm pretty sure it does but I will need to check with UTA / in the field.

> 2. What are the names of some bus stops? If they're named by the street
> they're on and the cross-street it should be possible to construct a name
> from the shapefile for each stop. (e.g. northbound far side 1st ave and
> main
> street)

They don't have official names that I know of, the stops are usually just
called by cross street (CROSSSTREET) or by address (ADDRESS + ' ' +
ADDSUFFIX + ' ' + ONSTREET) if there isn't one.

> 3. The shapefile appears to have address data for each stop. Should the
> addr:*=* information be added?

Wouldn't hurt if it's accurate. If you're going to put in an address
reference with house number accuracy, you'd expect (as a user) that that
house number is *right there*. If that's not the case I'd say don't bother.

I also noticed a couple of other things when looking at the data.
> The spatial accuracy is decent. Some stops are a few meters off and on the
> road, not the sidewalk, but they're all near the shelters that I can see.

Yes they are very proud of that and did not fail to mention that it's
spatially accurate data.

> Conflating this data with the existing data will be the hard part. If there
> hasn't been too much manual mapping of bus stops this could be done by
> hand.
> If there has, then you need to look at how to conflate the data.

No there has not been too much bus stop mapping. 202 stops (train/tram/bus
altogether) in Utah. See

martijn van exel
geospatial omnivore
1109 1st ave #2
salt lake city, ut 84103
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20120416/21cf7166/attachment.html>

More information about the Talk-us mailing list