I am attempting to calculate some measure of connectivity between a set of interconnected polygons. One idea I have is to randomly generate a number of points (say 1000) within these polygons then try to measure the average distance between each point and all other points. The more connected the polygons, the lower the average point-to-point distance should be.
The main issue is that each point-to-point distance should be the path distance ONLY within polygons, rather than the euclidean distance. I'm aware this is probably a network problem, but I don't have any experience with the network analyst extension, so I'm not sure how to do this.