A map edge is a modeling choice

Visionscarto’s “Planar vs. Spherical Voronoi” notebook on Observable compares airport Voronoi cells under several map projections. The useful part is not only the finished map: it makes the trade-off visible. A planar calculation is fast, but distortion around map edges and the antimeridian can become part of the result. When geographic accuracy matters, the notebook points toward spherical calculation with d3-geo-voronoi.

I like how the example keeps speed, projection, and precision in the same frame instead of treating the visualization as decoration.

Original notebook and code by Visionscarto:

https://observablehq.com/@visionscarto/planar-vs-spherical-voronoi