This example demonstrates a force-directed graph hierarchy on top of an email communication network, enabling users to drill up and down between aggregate community clusters and individual nodes. It first applies random-walk-based node embedding and k-means clustering to group nodes by community, then visualizes both the top-level community relationships and the detailed subgraphs within each cluster. The interactive drill-up and drill-down navigation is implemented using D3’s forceSimulation for layout, zoom behavior for pan and zoom, and SVG rendering. Color-coded nodes and edges distinguish clusters, with data parsed from a SNAP-format edge list and department labels.