[R-sig-Geo] How to deal with data points having the same lat long position?

Christopher Swingley cswingle at swingleydev.com
Fri Sep 5 21:29:58 CEST 2014


On Fri, Sep 5, 2014 at 10:29 AM, José Luis Rotundo <jrotundo at unr.edu.ar> wrote:
> I am working with a database of farm yields across USA and the objective is
> to map soybean yields. To preserve the identity of the farmers, they only
> provided the zipcode of the farm. Then I attached the zip code lat and
> long. The problem I have is that several farms belongs to the same zipcode
> and therefore they have the same lat and long. Having this repeated
> locations troubles the construction of the variogram.

Wouldn't it be better to attach the sum of the soybean yields assigned
to a zip code to a vector layer containing zip code boundaries?  I can
imagine that if you assign values to a single point (even jittered
when there are duplicates for single zip code) per zip code, that you
won't wind up with a very accurate representation of the spatial
pattern because the area each zip code covers will be different.


Christopher Swingley
Fairbanks, Alaska
cswingle at swingleydev.com

More information about the R-sig-Geo mailing list