[R-sig-Geo] Spatio-Temporal Models

Downey, Patrick PDowney at urban.org
Fri Apr 1 14:49:09 CEST 2011


I'm looking to estimate some spatio-temporal regressions controlling for
spatial and temporal autocorrelation and with a few independent variables.
The package {Stem} has exactly what I want, except my data is measured in
polygons, not at points.

Is there a package which can estimate these models when your data is not
point data (such as county-level population)?

I am under the impression that my estimates will be biased if I assign each
county's value to the coordinates at the centroid of the county and proceed
imagining that I have point data. Is this correct?

Thanks in advance for any guidance.


More information about the R-sig-Geo mailing list