I use ArcGIS desktop and used Optimized cluster routines to create the several measures.
Then I used these measures as dependent variables in GWR modeling.
My questions are
What are the pitfalls of this?
How do I interpret the results?
Is there some literature on this? With worked examples?
The data are 845 polygons of electoral and socio-economic measures for Germany 1919-1933. If others are working on GIS applied to historical quantitative data, I'm interested in corresponding.