Patient Stratification Clustering Simulator (2D)

This is the 2D companion to the 3D biomarker-space clustering sim: the same real k-means stratification, computed independently in a flat two-biomarker plane. Synthetic patients are placed by gene-expression score and protein biomarker level, and the assign-then-average Lloyd's-algorithm loop runs live in front of you — watch centroids converge as groups tighten around shared biomarker profiles, track the inertia chart fall every iteration, and click anywhere in the plane to drop in a new patient and see which therapy cohort they get triaged into, painted directly onto a live Voronoi decision map.