Graph Layout Algorithms

Interactive guide to the algorithms that arrange nodes on a canvas — from classic force-directed physics to modern spectral embeddings. Every demo is live and tunable.

Loading demos...

What Popular Tools Use

Layout algorithms used by well-known graph visualization and knowledge management tools.

ToolPrimary LayoutDetails
ObsidianForce-directedCustom implementation with 4 tunable force parameters
GephiForceAtlas2Also: FR, Yifan Hu, OpenOrd (1M+ nodes), circular, hierarchical
Neo4j BloomGPU force-basedTwo variants: compact and spread. Also hierarchical
Roam ResearchForce-directedd3-style graph view. Canvas features via tldraw
LogseqForce-directedSimilar to Obsidian. Whiteboard via tldraw
yEdFull suiteHierarchical, organic, orthogonal, circular, tree, balloon, radial, and more
draw.ioMultipleFlow, tree, radial, organic, circle, parallels, org chart

Best Practices by Scale

Different graph sizes need different strategies. What works for 50 nodes breaks at 5,000.

< 100 nodes

Any algorithm works. FR or KK for quality, dagre for hierarchy.

100 – 1,000

ForceAtlas2, fCoSE, or dagre. Memoize layouts.

1K – 100K

Multi-level (Yifan Hu, FM3). Barnes-Hut required. Node aggregation.

100K+

GPU-accelerated (cosmos.gl), OpenOrd. Density rendering, not individual nodes.