Fantastic work.<p>The final map for the US election in 2016 is both the most accurate and the easiest to understand I've seen so far:<p><a href="https://stemlounge.com/content/images/2019/10/muddy_america_2016_static-1.png" rel="nofollow">https://stemlounge.com/content/images/2019/10/muddy_america_...</a><p>In this map, saturation and lightness indicate vote density and color indicates the winning party, in each county.<p>IMHO, this is the kind of map that should be used by <i>every</i> media outlet to show election results.
I wonder how much difference it would make if vote "density" was used to determine the saturation, rather than total number of votes.<p>By my thinking, a geographically large county with 500,000 votes appears much more significant than a smaller county with the same number of votes in this map, and adjusting for density could potentially correct that?
A red-and-blue map is more like a brand logo for election news. The thumbnail for the Facebook or Twitter story. The saturated red and blue colors have an almost astrological meaning to people, it's got nothing to do with information.<p>Maps people struggle with this, they're always using maps to try to visualize a piece of data when almost always a short table would be better.<p>Data graphics people are themselves a subset of a family of wonks that spend 50% of their day rehashing the same tired stories, and the other 50% lamenting how innumerate people are.<p>Did you ever consider that maybe the reason the maps are stupid is because they're stupid as a whole, not because there's something wrong with the reader or the designer?
If you are going to insist on showing counties, why not also show boroughs of Alaska, and at their realistic sizes? If you do this the usual "OMG red land" depiction of 2016 gets flipped on its head, because the vast and empty regions of Alaska voted for Clinton.<p>If you want to argue against showing the boroughs of Alaska because nobody lives there, then we need to talk about all of those empty divisions of the lower 48, too.
This is cool.<p>I suspect that the value (white to dark scale) should be proportionate to votes/square km (or even better votes/pixel, since the projection isn't perfect at preserving area) instead of just number of votes though. In the current formula a huge county with 1000 is as dark as a tiny county with 1000 votes, but is visually much larger.
I had used similar data to speculate on the minimum number of people needed to move from the cities to the various counties in order to make the whole country a single party, in theory. A reality is that one party has the numbers of people necessary while the other does not.<p>I was curious to how much that would cost so it would be clear whether a super PAC could fund it. It would involve housing people long enough to be eligible to be registered in that state.<p>But then the pandemic happened and people did it on their own. We'll see!
It's a nice attempt, but capping the county population at 60000 means the map is effectively ignoring the majority of the population (if you consider the number of "ignored people" to be the sum over all counties of the county's actual population minus 60000)
Really nice work, but also quite misleading. The "real" map isn't interesting enough, so 97% of votes are excluded from the most populous liberal counties using an artificially low ceiling.