Where should Agent Architects meet?

A k‑means atlas of 476 community members — five theoretical meetup networks, drawn from one map.

1. Pulled 476 member coordinates from Skool's member‑map endpoint — the same data feeding the dots on the live community map.
2. Snapped each member to their nearest major metro (city pop ≥ 500k). A handful of US metros that fall below the threshold but have outsized metro areas — Miami, Tampa, Orlando, Salt Lake City, St. Louis, etc. — were force‑included so people don't snap to a different country.
3. Ran k‑means clustering on raw lat/lng for k = 10, 15, 20, 25, 30. Toggle Activity at the bottom to weight each member by their all‑time Skool points — centroids and hub size shift toward heavy contributors.
4. For each cluster, picked the host city = the metro where the most cluster members already live (tie‑break by global metro size).
5. All distances are great‑circle (haversine) from member home to host metro — not driving distance.