[Imports] [Imports-us] SNAP Retailer Locator
Yury Yatsynovich
yury.yatsynovich at gmail.com
Sat Oct 24 14:26:47 UTC 2020
Thanks for all your feedback!
As to Femont, CA -- indeed, I've started uploading the results, but, as one
could have noticed, it is done at a pace of ~1-2K buildings per day, after
manual edits to location/shape of each building, so the import is mostly
manual, not automated. With this pace it might be completed within 1-2
months. I'll update the documentation appropriately after it is finished.
True, sometimes errors like a temp tag "DONE=Y" get into the uploaded data
-- my apologies for this! By now I've already removed this temp tag and
will try not to make the same mistake again.
As to the retail locator -- the points in the SNAP data are, indeed,
located next to corresponding streets rather than on actual buildings. The
discrepancy between the SNAP points and the actual location of stores isn't
too large (10-30 meters for the points near my location that I checked),
but still not good enough for automatic import. I would do the import (if
the license permits) for MA by matching addresses of imported points to
addresses of OSM buildings (as MA has addresses added to many buildings),
rather than using only lat-lon of the SNAP points.
Thanks again, I appreciate your time and attention!
Best,
Yury
On Sat, Oct 24, 2020, 6:12 AM Rory McCann <rory at technomancy.org> wrote:
> Firstly, confirm the licence is OK. In addition, the data is from 2014
> (right?), are you sure there haven't been a lot of changes since then?
>
> I looked at the data in JOSM, It's about ~250,000 points, so this could be
> a useful resource. I don't know US culture, but I think SNAP is a type of
> social welfare from the state, right? It would be nice to be able to tag
> this service in OSM somehow.
>
> However, I looked at a randomn 20 points, with ESRI Clarity imagery and
> the position is all in the middle of roads, not on builds or other things
> that look like shops. So just uploading the 250k points to OSM is not
> possible. It would result in many OSM nodes that are in the wrong place.
>
> Rather than bulk uploading this CSV file, perhaps there's a way you couild
> try to semi-manually merge in to OSM, or to use it to detect missing data?
>
> On Sat, 24 Oct 2020, at 5:13 AM, Yury Yatsynovich wrote:
> > Greetings!
> > The US Department of Agriculture has the following data on retail
> > stores (https://www.fns.usda.gov/snap/retailer-locator), which is
> > pretty rich (244K points, lat, lon, names with ref and addresses). I've
> > just sent them an email asking about any license constraints on these
> > data (though, the description page
> > (https://snaped.fns.usda.gov/library/materials/snap-retailer-locator)
> > says "Free Material: Yes").
> > Any thoughts on the possibility of importing these data into OSM?
> > --
> > Yury Yatsynovich
> > _______________________________________________
> > Imports-us mailing list
> > Imports-us at openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/imports-us
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/imports/attachments/20201024/7d6430d4/attachment.htm>
More information about the Imports
mailing list