Permutation Feature Importance Lab — 2D Scatter View

Permutation feature importance answers a very concrete question about any already-trained model: how much worse do its predictions get if one input feature is shuffled so it no longer lines up with the sample it belongs to? This simulator fits a real linear-regression model to a synthetic dataset by ordinary least squares, renders every sample as a point on a 2D scatter plot positioned by its true feature value against the true target, and colors each point by the model's prediction error. Choosing a feature and running the permutation test shuffles that column, feeds the mismatched inputs back through the exact same fitted model, and recolors the plot live — a feature the model actually depends on turns the cloud red and visibly scrambles the x-vs-y trend, while an irrelevant one barely changes it. Averaging several independent shuffles turns that single noisy read into the stable importance ranking shown in the side panel. Drag to pan and scroll to zoom into any region of the scatter plot.