Note: If you are interested in learning more about how to build your own cartogram, please join my email list –>. I’ll be posting tutorials on how to build maps using this and other methods.
Judging by the amount of news coverage dedicated to looking at meaning of electoral maps, you’d would think we would have gotten better about it by now. Take this example map from the 2008 presidential campaign, where Barack Obama beat John McCain by over 12 percentage points (source).
By itself, maps tells us almost nothing. We know two counties voted for Obama more than they voted for McCain, but that alone does not answer the burning question at hand – who won the state’s 5 electoral votes?
Rounding up the usual suspects of map reading errors, the main culprit this time is population density. Because urban counties have a higher population density (and hence more votes) than rural ones, interpreting election results correctly requires the reader to be familiar with the population density of the state.
The second offender is the binary red vs. blue that we are so used to seeing. For presidential elections, where the electoral college chooses the outcome, this is an appropriate way to display election results. State elections, however, are determined by a popular vote, which means that it’s not the winner of a particular county that is significant, but the magnitude of that lead.
Fortunately for cartographers and election-watchers alike, a recent trend in reporting outcomes has emerged – the cartogram. By distorting a map relative to the number of votes an area received, not only do we get freaky-looking Frankenmaps, but we are actually showing the outcome that we care about in the end: the number of votes received.
So with that in mind, we present cartogram maps showing the outcomes of two of the most significant statewide ballot initiatives of November 2013.
New Jersey enacts cost of living-chained minimum wages with Public Question 2
With deep support in Essex, Hudson and Camden counties and more moderated support in the rest of the state, this cartogram clearly illustrates the voters’ approval by 20 points . of Public Question 2.
Since New Jersey is so uniformly dense, this cartogram looks surprisingly similar to regular county maps of New Jersey. If you came to this post hoping to see distortions, this next map is for you.
Washington state rejects GMO labeling with Initiative 522
Far closer result was a ballot initiative in Washington state. Had Initiative 522 passed, it would have required foods that were made with genetically engineered ingredients to be labeled. It failed by approximately 10 percentage points, but where did the votes fall?
The cartogram reveals two main results. First is the sheer dominance of King county over all other counties, as it makes up about 30% of the statewide electorate. Second, there was very strong disapproval of the initiative in the south and western counties, which despite their relatively smaller electoral might, was enough to overshadow the more closely split north and eastern counties.
The maps of the future
Maps exist to share information, so why do so many of the maps we look at fail to clearly display the essential information? Simply shifting from the red/blue dichotomy to gradients for popular vote-decided elections is an excellent start. For best mapping clarity, however, we really should be demanding from our news media cartograms like these on election night.
Can you find any maps (electoral or otherwise) that do a particularly poor job at showing their key information? Drop a link in the comment box below for others to see and learn from.
And if you have more general feedback, send me an email at firstname.lastname@example.org. I respond to each email I get.