Gerrymandering
We report several measures of compactness and partisan fairness. They often disagree, and no single number proves a map is gerrymandered — read them together and with the caveats below. Metrics are computed offline from Census TIGER district geometry. Source: TIGER/Line.
Partisan fairness by state
Partisan-fairness measures (efficiency gap, mean–median) need each district's election result on the current (post-2020) boundaries. This is now wired to our election-results feed and computes a state's metrics once all its U.S. House districts have reported certified results — until then we show nothing rather than a partial or placeholder number. Compactness below is already computed from real district geometry.
Least compact districts
- AK-00 Polsby-Popper 0.01 · Reock 0.13 · convex-hull 0.36
- NC-03 Polsby-Popper 0.03 · Reock 0.28 · convex-hull 0.47
- LA-01 Polsby-Popper 0.03 · Reock 0.32 · convex-hull 0.45
Methodology
Compactness: Polsby-Popper (4π·area/perimeter²), Reock (area / min-bounding-circle),
and convex-hull ratio — each 1.0 at most compact. Computed on an equal-area projection
(US Albers, EPSG:5070), which is tuned for the contiguous states, so Alaska, Hawaii, and
territories are approximate. Fairness: efficiency gap (wasted
votes) and mean–median (mean minus median district vote share). Boundaries are
versioned by as_of to track court-ordered changes. Ensemble/simulation
analysis is planned for a later release.