Skip to content

Interactive 分離

The segregated city

Drawn as Mexico City's census blocks — its AGEBs — this is Schelling's model: nobody demands a homogeneous neighborhood, each resident only wants a few neighbors like them, and still the city sorts itself into large segregated districts.

Urban pattern laboratory · 01

Schematic AGEB topology, Ciudad de México

Running
Group A Group B Vacant Boundary: INEGI MG 2020 · simulated cells

Live census

One move can alter an entire boundary.

Unhappy
0
Segregation
50%
Step
0

The alcaldía boundary is official; the cells and residents are simulated.

What you are looking at

The map

Charcoal and red residents on an irregular mosaic of blocks, like the census blocks (AGEBs) that tile Mexico City, with some left vacant. This is Thomas Schelling's 1971 model — the simplest machine that still produces segregated districts.

The one rule

Each resident is content only if at least n of their immediate neighbors look like them; below that they move to a vacant lot. That is the entire model — no group dislike, no landlord, no policy.

The surprise

Set the threshold to a mild 3 and run it. Weak, tolerant preferences compound across many small moves into a city that is almost fully separated. The Segregation readout climbs past 90% while nobody ever asked for that.

Things worth trying

  • Threshold 3, then run. Mild preference, hard separation. Watch the Unhappy counter fall as the board sorts itself.
  • Lower it to 1. Even "I just want one neighbor like me" is enough to start clustering — segregation does not require prejudice, only a mild comfort with sameness.
  • Crank it to 8. Everyone demands all eight neighbors look like them. The city never settles — it churns forever, because perfect homogeneity is unsatisfiable wherever a vacant lot exists.
  • Change the vacancy rate. More vacant lots let the city sort faster and more completely; fewer vacant lots freeze more mixing in place.
  • Pause and step by hand. Each step moves every unhappy resident once; watch the boundary between the two groups sharpen.
Teaching