Dasymetric disaggregation is an established approach for generating small-area population estimates while initially having that information on broader areas. It incorporates ancillary data to facilitate the areal interpolation process. A lot of work is done in the area of spatial population modelling. I wonder why I didn't found anything in R on it. Does anybody know if methods are available in R?
I've only read a little about small-area population estimates (Coulombe, 2009) and the interpolation was based on a linear model. This wasn't count data, but poverty. My spatial statistics professor, Daniel Griffith and Y. Chun (2013) recently published a book that includes a chapter on spatial interpolation using the spdep and gstat library in R. This covers kriging, co-kriging, EM, and spatial filtering. They developed an experimental spatial filtering method with eigenvectors that can be used with a Poisson regression. Perhaps you could adapt these for your small-area estimates.
I think you can try using the R function disaggregate. Hope it helps!
Here you can find both the methodology and the R code to implement it:
http://gsars.org/wp-content/uploads/2015/09/TR-Spatial-Disaggregation-and-Small-Area-Estimation-210915.pdf
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R