Random Graph Generator

What it does

The Random Graph Generator produces edge lists for five graph structures: random graphs (with a configurable edge probability), spanning trees, directed acyclic graphs (DAGs), weighted graphs, and hierarchies — up to 100 nodes. Output is a deterministic from–to edge list, optionally weighted, exportable as CSV, JSON, NDJSON, or TSV for import into any graph tool.

Common use cases

• Graph algorithm testing — BFS, shortest path, topological sort, and MST implementations verified against structures with known properties (trees are connected and acyclic; DAGs have no cycles).
• Network simulation — synthetic topologies for routing and epidemic-spread models.
• Visualization development — edge lists for force-directed layouts and dependency diagrams (presets for a small tree, medium DAG, and weighted graph).
• Database testing — graph-shaped data for Neo4j-style stores and recursive SQL queries.

Settings

• Graph Type — random, spanning tree, DAG, weighted, or hierarchy.
• Number of Nodes — 2 to 100.
• Edge Probability — 0 to 1 for random graphs: sparse networks near 0.05, dense near 0.5.
• Add Weights — attach weights for weighted-algorithm testing.
• Seed — identical seed + settings = identical edges.

Privacy note

Graphs are generated locally in your browser and never uploaded — pure synthetic structure, representing no real network or organization.

FAQ

What format is the output? An edge list — one from–to pair (plus optional weight) per row. It imports directly into NetworkX (read_edgelist), Gephi, and most graph libraries.

Are the structural guarantees real? Yes — trees come out connected with exactly n−1 edges, DAG edges only point “forward,” so algorithms with preconditions get valid input. Good material for property-based tests.

How do I pick an edge probability? Around ln(n)/n is the connectivity threshold for random graphs — below it expect disconnected components (also worth testing!), well above it graphs get dense quickly.