Partisan gerrymandering is generally considered to be an obstructor to true representative democracy. The efficiency gap metric is one way to measure the extent to which a particular district map benefits one party or the other in a bipartisan system: it measures the partisan difference in wasted votes (across the state) as a proportion of total votes cast, where a wasted vote is any vote for a losing candidate, or any votes more than floor(total_votes / 2 + 1) for the winning candidate.

This script applies the efficiency gap measure to Virginia's General Assembly and US Representative elections over the last 60 years, using data pulled from http://historical.elections.virginia.gov/.

For simplicity, votes cast for candidates other than Republican and Democratic are ignored. The efficiency gap works on the district map as a whole, rather than on any single district. To get a sense of the general balance of the district map, this script also sums across all elections for a particular office held within the decade for which the district map persisted.

While historically the district maps have varied in their partisan favour, in the last decade the VA and US House district maps have heavily favoured the Republican party - they have allowed the conversion of votes into seats at a rate nearly 20% higher than the Democrats

• At time of writing, the 2017 elections' results are not finalsed. The efficiency gap is particularly sensitive to results in marginal seats, so this election is ignored for now.

• The efficiency gap works on votes, and as such can say nothing about factors influencing those who did not vote at all. Uncontested seats have been ignored in this analysis for that reason, but low turnout by opposition voters in safe but technically contested seats is a potential confound.

• Python >=3.6

pip install -r requirements.txt

python3 -m va_efficiency path/to/figure.png

Path defaults to figure.png