US Regions

In my last post I demonstrated an algorithm that splits areas of the United States (lower 48) into subregions based on US Census population and boundary data. A main feature of that is using k-medoids to cluster adjacent census geographies, where adjacency was determined by a Delaunay triangulation of polygon centroids rather than by physical adjacency. Additionally, edge weights for the k-medoid calculation were re-weighted so that edge weights increased proportionally with real-world distance, but decreased quadratically as the populations of the linked nodes increase. This weighting is designed to capture the real-world effect of lower distance increasing the likelihood of real-world connections. Though the previous post focused on the counties of Long Island, this one will focus on actually separating the whole US into regions.

I used approximately the same algorithm as I had used in my previous Long Island post, but I did tweak the algorithm a bit after rethinking some assumptions I had made before. The inputs of this run of my algorithm were individual counties or county-equivalents of the lower 48 states. Additionally, I broke up a few regions at my own discretion because I thought they were too large: all of California except the very northern part had previously been in one region together with Phoenix, AZ. Most of Florida had also been included with most of South Carolina and all of Coastal Georgia. I decided to further break up those regions into two each, which are included in the map of all regions below.

In all I created 42 regions of the US based on my algorithm. Below I am grouping them roughly by their incorporation of their territories into the The United States or their thirteen predecessor colonies; I’ve tried to name the regions based on the two major cities within their limits. In future posts I plan to go through these 42 clustered regions in order.