Benchmarks of how well different two dimensional spaces work for clustering algorithms

This may be useful for guiding anyone optimizing using clustering for data science or machine learning problems. It also could be used to make cool animations.

Benchmarking is done by putting different numbers of points in the space, letting them repel each other for long enough to achieve equilibrium, and measuring how close they get to everything being equidistant.

To get metrics run from the command line.

There’s some highly opinionated interpretation and commentary in Commentary.txt


View Github