[Imports] [Imports-us] Address Data Import for Fulton County, Georgia

Saikrishna Arcot saiarcot895 at gmail.com
Tue Jan 7 22:50:46 UTC 2014


Serge,

On 01/07/2014 05:08 PM, Serge Wroclawski wrote:
> Saikrishna,
>
> I understand your enthusiasm for wanting to have address points for
> Fulton County, Georgia, but before you start uploading, we need to
> examine a few issues:
>
> 1. What is the license for this data?
>
> I was unable to download the data that was linked to from the wiki
> page. You have a statement saying that's it's OSM compatible, but the
> specifics are important- and in this case the county seems to require
> registration. They can do so, but then you (or someone else) need to
> legally be able to host the data- we don't want to require OSMers to
> sign up to evaluate this data.
>
The data is in the public domain. I've edited the wiki page to specify this.

I can upload the data elsewhere and make it available for download.
> 2. Address Points
>
> The second concern I have is that this is just a plain raw address
> dataset. We've discussed those before and I think it's safe to say
> that they're somewhat controversial. They're better than nothing, but
> often very difficult to work with, and when we have building
> footprints, there's going to be enormous work to conflate them.
>
> It would be much better if the conflation step was done before upload.
> Do you think it's possible to get building footprint data from the
> government?
>
I think I saw another data set that had building footprint data, but I
need to see what the data looks like and if the licensing is fine.
> 3. The Conversion
>
>  There are tags in the dataset which do not make any sense, such as
> STATUS=A and FILEVER, none of the tags have been lowercased, none of
> the addresses have been name expanded. All of this needs to be cleaned
> up and corrected before the data could be added to OSM.
>
Those tags are in the dataset itself; those exact tags won't be present
in OSM. The tags that will be used in OSM, as outlined in the Conversion
section, are addr:housenumber, addr:street, and possibly addr:city. The
casing and the name expansion will be done in the conversion program.
> 4. There's no mention in your wiki page about updates and any update
> process
>
I was initially under the impression that the addresses didn't have an
ID that could be used in updates. Looking at the data again, there is an
immutable GEO_OID field that represents a unique, internal ID. This,
perhaps, could be used for updates.

That being said, I was told that later versions of the address data are
under different licenses and so might not be possible to import into
OSM. (As it is, this is the latest version on the clearinghouse website).
> You have a good start here, but you have a lot to do before the import
> is ready to officially propose.
>
> - Serge

Saikrishna Arcot




More information about the Imports mailing list