Everyone is making UK election maps at the moment. Most of these are choropleth maps (coloured polygons), in which much of the UK (SNP revolution aside) appears as a sea of Conservative blue interspersed with clusters of red and the odd other colour. For example:

electionforecast.co.uk

There seems to be a growing critique in academia and data science questioning the appropriateness of the choropleth for political and sociological mapping because it misrepresents quantities with highly varied densities. Some alternative methods are various flavours of cartogram, of which an increasingly popular choice seems to be the ‘heatmap’ style* (where zones are gridded) as it can provide for a legible and elegant graphic. So to make myself useful I’ve developed shapefiles for UK constituencies using this method: box polygons and centroid points. I haven’t added Northern Ireland constituencies yet, but might on request.

This is the result, plotting the data from the same source as the map above. Boxes are coloured by most probable winning party, with intensity reflecting strength of probability.

gb_election_heatmap_cartogram

Methodology

The method uses Hungarian algorithm to allocate the actual constituency centroids to a grid with optimal efficiency (ie. minimising total displacement). This graphic illustrates the process.

* not all agree with this definition

Tagged with →  
Share →