[R] (no subject)

adrian at maths.uwa.edu.au adrian at maths.uwa.edu.au
Wed Aug 15 04:39:00 CEST 2007

"Lawrence D. Brenninkmeyer" <ldb at outsourcethinking.com> writes:

> I am trying to find a way to diffuse GIS data on a European map. I have a
> dataset consisting of particular locations scattered across Europe,
> along with magnitude and value information.

The dataset is a spatial point pattern, where the locations are points,
and the value or magnitude is a `mark' attached to the point.

You can use the package 'spatstat' to smooth the values.

# extract coords
         xy <- geocode[,c("LONG","LAT")]
# make an enclosing rectangle
         Box <- ripras(xy, shape="rectangle")
# create point pattern
         X <- as.ppp(xy, Box) %mark% (geocode$VALUE)
# visual check
# smooth using density.ppp
         smoothVP <- density(X, weights = X$marks)
         smoothP <- density(X)
         smoothV <- eval.im(smoothVP/smoothV)
# inspect

The result 'smoothV' is a pixel image (object of class "im") that you can
manipulate, threshold, take histograms, etc

# plot the region where smoothed value exceeds 42
         plot(solutionset(smoothV > 42))

More information about the R-help mailing list