Spatial Randomness Evaluation in R: Monte Carlo Test

May 14, 2012
By

(This article was first published on Misanthrope's Thoughts, and kindly contributed to R-bloggers)

This post is a some kind of reply to this one.

So our goal is to determine whether our point process is random or not. We will use R and spatstat package in particular. Spatstat provides a very handy function for this, that uses K-function combined with Monte Carlo tests. I will spear you from burbling  about theory behind it - the necessary links were already provided. Lets get directly to action.

In this example I will test data about location of my "favourite" illegal dumps in St. Petersburg and Leningrad region.

# we will need: 
library(maptools) 
library(rgdal) 
library(spatstat)

# import data for analysis

S <- readShapePoints("custom_path/dump_centroids.shp", proj4string= CRS("+proj=tmerc +lat_0=0 +lon_0=33 +k=1 +x_0=6500000 +y_0=0 +ellps=krass +towgs84=23.92,-141.27,-80.9,-0,0.35,0.82,-0.12 +units=m +no_defs"))

SP <- as(S, "SpatialPoints")
P <- as(SP, "ppp")

# perform the test itself with a 100 simulations

E <- envelope(P, Kest, nsim = 100)
plot(E, main = NULL)


And here is what we've got in the end, a fancy graph, which demonstrates that our data (Kobs) significantly deviates from a random process (Ktheo):



To leave a comment for the author, please follow the link and comment on his blog: Misanthrope's Thoughts.

R-bloggers.com offers daily e-mail updates about R news and tutorials on topics such as: visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...



If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.