2

I have previously asked this question but perhaps was using the wrong terminology. I would like to run some type of cell declustering on a spatial dataset. Is there a tool for doing this in Arc? Alternately, if I overlay a grid over my points, how would I randomly sample 1 point per grid cell - this would be one way of generating a declustered set of points. I have geostatistical analyst but I don't see the option to spit out a declustered dataset.

OLD QUESTION: In ArcGIS 10.2, is there a simple tool/process for declustering a spatial dataset via Spatial Analyst package? My data are spatially clustered in some areas. I want to filter some locations out such that I end up with an equal density of points throughout. (Kriging achieves this by weighting data values differently depending on the density of point values around the center point.)

I've looked a the resource notes and they suggest using the Geostatistical Analyst Wizard, which I don't have.

  • 2
    Spatial Analyst is more for raster datasets. Unfortunately, you'll need Geostatistical Analyst if you want to use ArcGIS. There are probably open source options for QGIS or GRASS. – Fezter Nov 6 '15 at 2:53
  • Can you provide some more detail? It's pretty vague what you're trying to do. – Leo Nov 6 '15 at 19:49
  • @LeonB I've updated the question. – val Nov 6 '15 at 22:36
1

I was able to solve my problem using these steps:

  1. contruct a fishnet over my sample locations (using some user-defined size)
  2. convert fishnet cell polygons to points at the center of each cell
  3. join my center points to the nearest sample point.

This effectively declustered my data, but I would be interested to hear from others that had faster methods or the ability to use different declustering methods (nearest neighbour, random sampling, average of points, etc.) enter image description here

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.