When looking at data like this, I'm always reminded of Schelling's model of residential segregation, which suggests that even a small preference for living near

When looking at data like this, I'm always reminded of Schelling's model of residential segregation, which suggests that even a small preference for living near like people leads to high segregation in the global system. Example: in this simulation, agents are willing to live in environments where only 30% of neighbors are like them: https://www.youtube.com/watch?v=dnffIS2EJ30 https://www.youtube.com/watch?v=dnffIS2EJ30 Another good explanation of it: https://www.youtube.com/watch?v=dFl3Cfw12bo https://www.youtube.com/watch?v=dFl3Cfw12bo I wonder if anyone has done a "reverse" model, taking in census data and trying to estimate a rough "Schelling threshold" -- like if all the dots started out randomly placed on the US, what percentage of "tolerance" would lead to the clustering we have in reality.