Maps in election reporting
In several sections of How Charts Lie, Alberto Cairo discusses the often seen misuse of data presented on maps, especially electoral data. I strongly agree. A good example of this phenomenon can be seen from this image taken seen on the BBC coverage of the UK election:
We can see from the table that the Conservative party won 45% of the vote versus the Labour party winning 33%. However, it looks from the map that the Conservatives have almost all of England and Wales, while the Labour party have almost none. It seems that the SNP have the third of the vote versus their actual 4%!
This, of course, reflects the principle that hectares does not equal headcount! The very small red areas of the map are on larger urban areas such as London, Cardiff and Liverpool where there are higher populations. Scotland, where the SNP did so well, has a relatively small population and many of the large areas of land have a very sparse population.
One type of map that has been used to more accurately portray the spread of vote in an election is the cartogram. My personal favourite variant for this purpose is the use of hexagons, where each hexagon represents a roughly equal share of the voting population (or college votes) rather than land area. I first saw this from the New York Times coverage of the 2012 US Presidential election.
In fairness to the BBC, they have given users the option to see this type of chart on the website, and have been using the hexagon cartogram as a physical map that presenters can walk across. Each hexagon represents one constituency which is one seat in the parliament.
From this cartogram, we can see a more accurate visual of the scale of the Conservative victory. The urban areas that voted for Labour are clearer.
Well done BBC on adopting this approach. We should all strive to use data visualisation to inform rather than misinform!