Teaching simulators

Simulators

Interactive tools that open up the algorithms. Each simulator lets students control the steps one at a time, tune hyperparameters and see how each decision is made. Everything runs in the browser, with no server.

Family

Clustering

live

DBSCAN

Density-based clustering. Watch the algorithm classify each point as core, border or noise and grow clusters by contagion. Control eps and MinPts.

open simulator

live

Hierarchical Clustering

Bottom-up agglomeration. Watch the scatter plot color merges while the dendrogram grows bar by bar on the right. Compare Ward, single, complete and average.

open simulator

live

Gaussian Mixture Models

EM step by step, with separate E-step and M-step. Points recolor by responsibility and ellipses fit the covariance until convergence.

open simulator

Family

Dimensionality reduction

live

PCA

Follow principal component analysis step by step: centering, covariance, eigendecomposition and projection onto PC1/PC2 with the explained variance.

open simulator

live

UMAP

KNN, fuzzy union, embedding initialization and optimization. Original space and 2D embedding shown side by side.

open simulator