[R-sig-Geo] Clarification on Neighbours list

[Ronnie Babigumira]

Roger (as well as others who contacted me off the list), many thanks for this it helped clarify things.

Ronnie
Roger Bivand wrote:
> On Thu, 5 Jan 2006, Ronnie Babigumira wrote:
> 
>> Someone has pointed out that my question was not clear. With a clear danger of sounding really stupid, I will try to clarify
>>
>> Say I have a polygon, which I convert to raster format so each cell is n by n. For some reason I select a set of cells 
>> and have a neat rectangle such as below. To get a neighbor list for such a grid I could use cell2nb(3, 8) would be fine
>>
>> 1 1 1 0 0 1 0 1
>> 0 1 1 0 1 1 1 0
>> 1 1 0 0 1 1 1 1
>>
>> However say my grid came from the full map and looked something like this
>>    0 1 0 0 1
>> 0 1 1 0 1 1 1
>>      0 0 1 1 1
>>        0 0 1 0 1
>>    1 1 0 1 1 1
>>    1 0 0 1
>> 1 1 1 0 0
>> 0 1 1
>>
>> The number of rows is clear, however, what is the number of columns (and anyway, what R function do you use to get the 
>> dimensions of a SpatialGridDataFrame object.
>>
> 
> At an enjoyable workshop south of Leipzig in November, this came up. If 
> you have a SpatialGridDataFrame from a raster, with NA where there are no 
> values:
> 
> x <- read.asciigrid(system.file("external/test.ag", package="sp")[1])
> summary(x)
> xx <- as(x, "SpatialPixelsDataFrame")
> summary(xx)
> slot(xx, "grid")
> # find grid spacing, 40 distance units for rook neighbours
> # for queen, max distance about 57 should work
> d40_nb <- dnearneigh(coordinates(xx), 0, 40)
> d40_nb
> # convert to polygons
> xxsp <- as.SpatialPolygons.SpatialPixels(xx)
> # generate rook neighbours again - takes longer
> rook_nb <- poly2nb(xxsp, queen=FALSE)
> rook_nb
> # compare, alike apart from different IDs
> diffnb(rook_nb, d40_nb)
> 
> Hope this helps,
> 
> Roger
> 
>> Ronnie
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at stat.math.ethz.ch
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>