H3: Uber ridesharing uses hierarchical grid indexing with Hexagons

Uber open-sourced a library to hierarchically index all points on the earth using Hexagons:

Key insight: they had to choose between triangles, squares, and hexagons… and interestingly, tiling with either triangles or squares led to more neighbor tiles (squares have 9 neighbors).

Some fun pictures

Github source: https://github.com/uber/h3