The thing is, the maps all show the difference was the Suburban Times didn't take Sawant seriously, so they and Downtown didn't tar her with the War On Cars lie or the Downtown Crime Spree lie they used against McGinn.
Isn't this the flip image of the last set of red:blue states for the presidential campaign, where states that bordered large bodies of water generally were blue?
@7: But this isn't the same at all, because it is looking at percentage, not flat value. If it was a map showing # of Sawant voters, that would be a legitimate comparison, but it is not, and thus your comparison is inapt.
@12 of course i humbly disagree. if more than a few of those delineated regions match up (and many match up exactly) then any proper analysis should find a way to compensate ("subtract") out the population bias; which has not been done. as it is, any statistician worth her salt would tell you that very little can be concluded from the way it's been presented. the heat map xkcd complaint is valid and quite "apt" because you're deceptively tracking population as well as vote percent. (i don't see how it being a "flat value" or percent or relative rank or parts-per-million ratio would ever validate something invalid like this)
@13 The regions in your population density map (Census block groups) contain very different numbers of residents and are basically uniform in size. The regions in the map on this post (precincts) are basically uniform in population, and vary widely in size. That patterns of population density are consistent with patterns of Sawant's strongest precincts means that people who live in dense neighborhoods tended to vote for Sawant. Supporting one candidate vs another, and by how much, is not inherently related to population.
Isn't it a bit disingenuous to use Blue-Red for this map? Sure, Sawant is further left than Conlin, but Conlin isn't a Republican by any means. And technically the race was non-partisan. And in liberal Seattle making Conlin-voting districts red makes them seem evil...
A simple Green-yellow scheme would have been much more neutral and just as informative.
They will use it next time.
(e.g. http://buildthecity.files.wordpress.com/… ) so to a first approximation of pattern recognition one has to raise the statistician's lament: http://xkcd.com/1138/
A simple Green-yellow scheme would have been much more neutral and just as informative.
Best Seattle indeed.
And just because they are red doesn't mean conservative. Conlin wasn' exactly a big conservative no matter what slog would have you believe.