As mentioned in my previous post, Geographically-Weighted Regression (GWR) is an extremely effective regression model for spatial analysis, especially when there may be regional variance in the relationships between independent and dependent variables. GWR allows us to explore the local relations amongst a set of variables, and to examine the results spatially using ArcGIS; it can even produce raster coefficient surfaces which allow us to see any regional variation in the relationships of the parameters! We can see if the variables and residuals of the model are spatially dependent or spatially autocorrelated. Continue reading
