Using hexagon tilemaps for election outcomes

The November 3, 2020 presidential election in the United States has a clear winner: congratulations to President-Elect Joe Biden and Vice President-Elect Kamala Harris. As of November 15, 2020, Biden and Harris were projected to hold 306 electoral college votes, well over the 270 margin needed to win. Four years of hyperpartisan politics by the outgoing Trump administration have left the United States a bitterly divided country. Healing these wounds and restoring trust in the democratic institutions of the United States will be a tall order for the new administration. A large segment of the US electorate appears to have dissociated itself from reality and instead believes in whatever lies and conspiracy theories are dished up by Trump and his enablers and acolytes. With the outcome of the Senate elections hanging in the balance with two runoffs in Georgia, the path towards urgently-needed new policies in the United States looks increasingly difficult.

Today's blog is primarily about how to depict the 538 electoral college votes. Below is a hexagon tilemap where each hexagon represents one electoral college vote. A tilemap is a type of tessellated cartogram that morphs the spatial relationship of geographic units into a shape that depicts the thematic size of the geographic unit, such as population or GDP or electorcal college votes.

click here for PDF version of image
Source: Tilegrams by Pitch Interactive

The chart shown above is based on the tilemap generated by Pitch Interactive which they have made available on Github. I have adapted it with my own software to generate an SVG graphic for this page. The German newspaper Die Zeit also uses it to show the results of the US election (see "Die Ergebnisse").

Population in the United States is heavily concentrated in urban areas, and thus populous states that have a small land size are enlarged, while sparseley populated rural states are shrunk in size. Tilegrams can be created with squares as well, but hexagons are particularly good at approximating shape because they are "rounder" than squares.

R has a package for making tilegrams, tilegramsR, written by Bhaskar Karambelkar. Making hexagon tilemaps ("hexmaps") is not trivial, however. The underlying algorithm solves a linear assignment problem (LAP) with a distance matrix of hexagon locations and geographic centroids. Specifically, variations of the Hungarian algorithm are used to solve this assignment problem However, the map also needs to preserve topology—which state borders which other state—and give perhaps greater weight to the overall outline of the country than internal borders. Hexmaps should not have geographically inaccurate holes or other artifacts. In many instances, computer-generated results will need a little bit of manual fine-tuning to make the result visually pleasing.

Some final remarks on the US 2020 election from the perspective of a Canadian. The electoral system in the United States is highly deficient. The electoral college system for electing presidents is as outdated as it is biased; it locks in a permanent advantage for Republicans. For congressional districts, partisan gerrymandering is further distorting results. Independent non-partisan electoral commissions could easily fix this problem, as litigating these distortions has so far gone nowhere with the US Supremce Court. As electoral districting and voter registration falls upon individual states and the US Voting Rights Act of 1965 and was largely dismantled by the Shelby Count v Holder decision in 2013, it appears likely that voter suppression tactics will endure. Short of constitutional amendments (the likelihood of which appears close to zero), democracy in the United States will continue to suffer from partisan distortions of the electoral process.

Posted on Sunday, November 15, 2020 at 11:15 — #US | #Politics | #Software