Skip to content

package to query Google for latitude and longitude for a given ZIP code

5 messages · Dan Putler, Andrew Yee, Tom Sgouros +1 more

#
Andrew,

What you are proposing (a) violates Google's terms of service and (b) is 
pretty meaningless since a PO Box zip code will be located at a post 
office location, not the household's or firm's location, which I'm 
pretty certain is not the location that will matter conceptually for any 
model you develop based on this information.

Dan
On 03/08/2011 09:03 AM, Andrew Yee wrote:
#
Thanks.  That's interesting, with respect to Google's terms of
service, since I thought this kind of information was available
through the Google Geocoding API.

The reason why I'm pushing for the P.O. box ZIP code is that some
geographic information is better than none.

Thanks,
Andrew
On Tue, Mar 8, 2011 at 12:16 PM, Dan Putler <dan.putler at sauder.ubc.ca> wrote:
#
Andrew:

I don't know if it matters to your goal, but there are some zip codes that do not correspond to locations, and some correspond to locations best represented by a line, not an area.  The post office originally set them up to represent delivery routes or groups of routes, and most of them represent a place, but not all do.  I ran across a note about the limitations of zip codes as a representation of areas in the technical docs of some US Census data once.   I know of one zip code near me that contains a couple of buildings that aren't next to each other.  Some geographic information is better than none, but I don't know what you'd do about a situation like that.

 -Tom
On Mar 8, 2011, at 12:16 PM, Dan Putler wrote:

            
#
Read the terms of service agreement. You can use the service for the 
purposes of displaying information on a Google map, but not store the 
information or use it for non-display purposes. To use it in R, you will 
be using it for non-display purposes and you will likely store it. This 
topic has been discussed on this list before.

You could use the city, and take advantage of the U.S. Board on 
Geographic Names (public domain) data. See http://geonames.usgs.gov
On 03/08/2011 09:23 AM, Andrew Yee wrote: