Equal Person Open Districting Project

Equal Person Open Districting Project 2020 census block groups · apportionment seats · one rule

State

Loading state…

6 districts · about 706,209 people each Colours are graph-coloured so no two neighbours share one

Deviation target Seed Growth Smoothing

How these districts are drawn

Quota. The state's 2020 population divided by its seats. Oregon: about 706,000 people per district.
Seed. Start at the densest unassigned block group in the largest remaining region.
Grow. Add the neighbour that keeps the outline tightest per person added, until the quota is met. One district at a time, watched live above.
Repair. Regions walled off behind a finished district are merged into the neighbour with the most room. The people moved are counted, not hidden.
Smooth. Move boundary block groups between districts when it shortens both outlines and keeps both inside the population bound. Contiguity is checked with articulation points, so a district never splits.

Every number here is produced by the same code that draws the map. Population is 2020 census POP100 at block-group level; the quota is resident population over seats, which is the definition the block groups can actually sum to. The legal quota uses apportionment population, which adds federal employees stationed overseas — about 0.1% more per district, and invisible in any census block.

What this is not: a Voting Rights Act analysis, a partisan model, or a claim that a compact map is a fair one. Compactness is measured and reported as one property among several. The 119th-Congress comparison and block-level granularity are the next two releases.

The deviation floor is set by the largest block group in the state, not by the slider. The stats show both, because a bound the algorithm cannot reach is not a result.

This map needs JavaScript

The districts are drawn in your browser from 2020 census block groups, so there is nothing to show without it. What the map does:

Divide each state's resident population by its apportioned seats to get a quota.
Seed a district at the densest unassigned block group in the largest remaining region.
Grow it one neighbour at a time, always choosing the addition that keeps the outline tightest, until the quota is met.
Repair regions walled off behind a finished district, counting the people moved.
Smooth boundaries where a move shortens both outlines and keeps both districts inside the population bound.

Measured across the 49 states with maps: worst population deviation 0.00%–4.98%, with the floor set by the largest block group in each state (0.45%–4.98%). Compactness (Polsby-Popper) 0.12–0.77. Alaska has one seat and no map: it crosses the antimeridian, so a projected outline would be meaningless and there is no line to draw. This is not a Voting Rights Act analysis, not a partisan model, and not a claim that a compact map is a fair one.