Hands-On Geospatial Analysis with R and QGIS

Spatial segregation for a bivariate marked point pattern

For bivariate marked point pattern processes such as case and control events, a null hypothesis could be the probability of disease being the same in every place against the alternative of it not being the same in all places. The spatialkernel package allows us to conduct significance tests for spatial segregation. First, we'll need to choose a bandwidth for kernel smoothing. Using the spseg() function, we can get it in the following way:

bandwidth = spseg(marked_ppp, h = seq(0, 100000, by = 50), opt = 1)

Now, we are looking for a bandwidth between 0 and 1,00,000 meters with a 50-meter step. Here opt = 1 tells spseg() to look for the best bandwidth. This returns a list and the best ...