[Geocoding] General Nominatim questions

Charles Randall crandall747 at gmail.com
Mon Oct 15 23:08:24 BST 2012


Thanks for the responses Sarah. I didn't understand address details fully
before, but on second read it makes sense. Thanks for all the info!

Charles.


On Mon, Oct 15, 2012 at 2:01 PM, Sarah Hoffmann <lonvia at denofr.de> wrote:

> Hi,
>
> On Mon, Oct 15, 2012 at 11:33:17AM -0700, Charles Randall wrote:
> > Hey all -
> > I've got a couple of random Nominatim questions for the group:
> >
> > a. Nominatim accuracy: A number of times, when I put in a specific house
> > address for Nominatim, I get back a lat long, which doesnt vary with
> street
> > numbers.
> > - I assume this is because of the data files used.
>
> This happens, when the house numbers are not in the database, then the
> coordinate of the street is returned.
>
> > - Do people do anything to make these more accurate?
>
> You can always survey house numbers and add them to the OSM database.
>
> > b. Address details option in URL:
> > I'm curious, if I set the nominatim query string to have the option
> > "addressdetails=1", what exactly is coming back in the data I see
> returned?
> > I feel like the different sets of place id's returned is the lat long
> > address details of the other street options that are listed on the main
> > screen when you do a query.
>
> I'm not sure I understand that question. What you get back with
> addressdatails
> is a list of the parts of the address together with their designation
> (street,
> city, county, country, etc.)
>
> > c. Are there particular xml files for US address data that some people
> find
> > are more accurate than others?
>
> Not sure how that relates to Nominatim. It's primary source is data from
> OSM,
> there are no other xml files. You'll find different excerpts in the
> internet
> but they all come from one source - OpenStreetMap.
>
> For the US, it can also be made to use Tiger house number data on top of
> the
> OSM data, which will most help with your problem from question 1. I can't
> really comment on the accuracy of this additional TIGER data. Maybe others
> know better.
>
> Sarah
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/geocoding/attachments/20121015/e86f9795/attachment.html>


More information about the Geocoding mailing list