I am working on a project looking at county boundaries for the state of California. I have 2 polygon datasets with this same geographic boundary, and these datasets, when mapped individually, visually appear to have a spatial correlation. Polygons are mapped according to numerical attribute values, which range widely on both datasets.

Wondering what the best method might be to test for a correlation between these datasets?

  • You may want to review previous Questions on Correlation using ArcGIS for Desktop here.
    – PolyGeo
    Commented Jan 22, 2015 at 23:34
  • I am using shape files in ArcMap 10.1. Commented Jan 23, 2015 at 1:02
  • If both shapefiles are of identical spatial features, you can skip most of the steps in FelixIP's answer. If they both share a unique ID field (FIPS, county name, etc.), you can simply join one to the other. It doesn't sound like this is a necessarily spatial issue; is that correct?
    – Tom
    Commented Dec 10, 2015 at 15:45
  • Why not run a simple spatial linear regression to see behavior and direction of each polygon in the relation.
    – Vlad
    Commented Feb 2, 2021 at 22:58

1 Answer 1

  1. Intersect 2 layers
  2. Add field [Times] and calculate it using 2ndNumeric*ShapeArea
  3. Summarise field [Times] and [ShapeArea] using first layer UniquPolygonId
  4. Divide [Sum_Times] by [Sum_Area], this will give you weighted second layer numeric
  5. Transfer result to 1st polygon table and check the correlation between 2 columns, i.e. 1st numeric and weighted one

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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