EWMA Outbreak Surveillance: The 2D Control-Chart Grid

Public-health machine learning rarely means a black-box neural network — much of it is disciplined statistical monitoring applied at scale. This simulator recreates a syndromic-surveillance early-warning system: twelve regions report daily case counts, and an exponentially weighted moving average (EWMA) control chart watches each region's smoothed statistic against a control limit derived from its own baseline noise. Inject a simulated outbreak with an adjustable growth rate, tune the EWMA weight and alarm threshold, and watch how fast — or how falsely — the detector reacts as the 2D region tiles rise and fall and the affected region's color crosses from calm cyan to alarm red, while the live chart on the right tracks the exact EWMA statistic against its control limit.