An interactive gerrymandering simulator that runs entirely in the browser. You paint congressional districts on a procedural hex-tile map and see how your choices move fairness metrics in real time.

Try it | Part of the a9l.im portfolio

Draw district boundaries on a randomly generated map populated by three political parties (Federalist, Farmer-Labor, Reform) with realistic urban, suburban, and rural demographic leans. The simulator scores your map on six fairness metrics (efficiency gap, partisan symmetry, competitive districts, compactness, contiguity, and majority-minority representation), so you can see firsthand how and why a given map is fair or unfair.

Two automated algorithms let you compare extremes: a pack-and-crack gerrymander that maximizes seats for a chosen party, and a Lloyd's relaxation fair draw that optimizes for population balance and compactness. A Monte Carlo election simulator stress-tests any plan against random vote swings.

• Gerrymandering tactics: packing and cracking, three-party dynamics where you can sacrifice one opposition to benefit another
• Efficiency gap: wasted-vote analysis across three parties, with the 7% warning threshold used in real court cases
• Partisan symmetry: whether a map treats parties equally when vote shares are swapped
• Compactness: Polsby-Popper score measuring how circular districts are
• Contiguity: whether all hexes in a district are physically connected
• Majority-minority districts: ensuring minority communities get adequate representation
• Monte Carlo simulation: running 50 to 500 elections with Gaussian vote swings to see which plans are resilient

Click or drag to paint hexes with the active district color. Right-click to erase. Use the numbered palette at the bottom to switch between 8 districts, or press keys 1-8. The sidebar updates fairness metrics as you draw, so you get immediate feedback on every change.

Toolbar buttons give you auto-fill (greedy fill to population target), automated gerrymander, fair draw, election simulation, and plan save, load, and export. Press ? for the full keyboard shortcut overlay.

cd path/to/a9lim.github.io && python -m http.server

Shared design system files load from the parent directory via absolute paths, so please serve from the repository root. There's no build step or dependencies.

Vanilla JS with no dependencies. SVG rendering with event delegation (no per-hex listeners). Seeded PRNG (Mulberry32) for reproducible maps, and seeds are stored in URL hashes for sharing. ES6 modules load directly without a build step.

AGPL-3.0