/spst

adventures in spatial statistics -- with c!

Primary LanguageC

spst: spatial autocorrelation, by force

  • for now, this only calculates the global moran's i for a shapefile
  • don't use epsg4326 (or any other unprojected crs). use something like 3785.
  • distances less than one are bumped up to one.
  • did i mention this only does idw?
  • and that distances less than one are bumped to one?
  • so you shouldn't use an unprojected lat/lon srs.
  • on my laptop, which is decent, it takes just under 7 seconds to build the idw and calculate global moran's i for 1000 features.
  • i know that's not very fast.
  • you should try doing 100000 features. or don't.
  • 10000 features takes about 12 mins. lolz.
  • there's also a small memory leak, but i'm not sure if it's something i'm doing wrong with geos or ogr or what. i'll get around to it.

spst file.shp colname