moranfast: Calculate Moran's I quickly with low memory footprint for Big Data.

This package calculates the Moran's I test statistic of spatial autocorrelation for point observations.

This package is still in development. Right now it just does one thing (calculates Moran's I).

moranfast is an improvement over any other package I know of for calculating Moran's I for two reasons:

1. It is memory efficient, because it calculates the distance matrix on-the-fly. It shouldnt take up much more memory than it takes to hold a dataframe of point observations in R.
2. It is fast, because it uses Rcpp. I found it calculated the Moran's I for 100,000 observations in under a minute.

This package drew significantly on the Moran.I function in the ape package.

Since this package is still in development, you have to install it with the devtools package:

library(devtools)
install_github('mcooper/moranfast')

library(moranfast)

ozone <- read.table("https://stats.idre.ucla.edu/stat/r/faq/ozone.csv", sep=",", header=T)
moranfast(ozone$Av8top, ozone$Lon, ozone$Lat)