Skip to content

Glossary · Interaction

Graph Clustering

Also known as: Node Clustering, Compound Nodes.

Grouping graph nodes by a shared attribute (department, region, type) and aggregating the edges between groups so users can collapse and expand whole subgraphs.

Real-world graphs often contain natural groupings — employees by department, services by region, processes by team. Graph clustering exposes these groupings in the visualization: nodes belonging to the same group are visually associated (often with a surrounding shape), and the edges between groups are aggregated into thicker summary edges.

Clustering serves two goals. First, it gives the viewer a high-level map: instead of 5,000 individual nodes, they see 12 clusters with weighted connections. Second, it enables drill-down — click a cluster to expand it and reveal its members, click again to collapse.

In TopoKit, the clustering API is `cluster.by(field)` with optional edge aggregation and collapsible groups. It works with any layout algorithm.

See also

Clustering (docs) Org Chart (use case)

References

Browse the full glossary or jump to the docs.