[Talk-us] Address improvement through imports?

Michal Migurski mike at stamen.com
Thu Nov 3 03:38:22 GMT 2011

On Nov 2, 2011, at 9:28 AM, Martijn van Exel wrote:

> On Wed, Nov 2, 2011 at 10:23 AM, Anthony <osm at inbox.org> wrote:
>> On Wed, Nov 2, 2011 at 12:19 PM, Martijn van Exel <m at rtijn.org> wrote:
>>> I'm all for not importing data where there's existing data people can
>>> use, but in the case of TIGER addresses you could actually make a
>>> point for importing: OSM could be a platform for improving that
>>> address data (like it should be for the GNIS points). 'All we need' is
>>> a suitable microtasking platform.
>> I can't speak for other locations, but here in Florida there is much
>> better address information available from the counties than that
>> available from TIGER.  So if you're going to build a suitable
>> "microtasking" platform for Florida, use the county data, not TIGER
>> :).
> Couldn't agree more - we have to make an effort to import the best
> address data available and that probably means looking at the local
> government level. That's why I made that page I started this thread
> with in the first place - to try and consolidate that effort.

I like this approach.

Ian and others in the thread have described address imports as a sort of "computer-assisted manual editing", and I think for US addressing it's a great approach. The gulf between urban and rural parts of the U.S. is wide, and if we don't take advantage of the excellent local government data available for these purposes there are going to be regions of the country that *never* get mapped.

I believe that we should develop an approach that is welcoming to local government representative who can help with the assisted importing, which may involve doing a few test imports in known places followed by the development of new tools or description of new procedures that newcomers to OSM who already steward local address databases can repeat for their own jurisdictions.

First step would probably be a plan for shitty TIGER data - what should these local government folks do in cases where their local OSM coverage is pure TIGER 2007 drek? Just to name a local example, Mill Valley CA is right across the bay from SF and composed of mostly import junk. Maybe "someone" (heh) could do a purpose-built fork of Potlatch designed especially for pulling in address info without displaying any other road data to eliminate confusion, for use by owners of address info who know they have good, high-precision coverage.


michal migurski- mike at stamen.com

More information about the Talk-us mailing list