Dimensionality Reduction Meets Network Science: Sensemaking on UMAP’s kNN Graph

This post introduces the application of standard graph algorithms (PageRank, k-core decomposition, clustering coefficient) to UMAP's internal kNN graph for data sensemaking. It demonstrates how these algorithms can identify representative data points, reveal dense core regions, and detect similar data neighborhoods, providing quantitative and qualitative evaluations on MNIST and Fashion MNIST datasets.