🧪 Data Science Model Evaluation Simulation

This simulation illustrates how to evaluate the performance of a binary classification model. Two synthetic populations of classifier scores — real positives and real negatives — are rendered as 3D histogram bars, and a moveable decision-threshold plane sweeps across them. Every bar that crosses the plane changes which quadrant of the confusion matrix it belongs to, and accuracy, precision, recall, F1 and the ROC curve all update live, so you can see directly why "best model" depends on which metric — and which threshold — you choose to evaluate it by.