I’m sure there are quite a few things that each of us would like to leave behind in 2016.
Bizarre presents from distant relatives, stressful deadlines at work, sporting disappointments. All these things represent the mundane lowlights of a year, which are best left on the forgotten side of January 1st.
But there is something else: a piece of data visualization which I would like to see consigned to the scrapheap of history.
I feel I’m not alone here. This visual offender of epic proportions crops up time and time again, whenever we have a presidential election. It tends to ruffle some feathers.
Fans of pretty red and blue color schemes and geographically-accurate depictions of our beloved U. S. of A. probably don’t mind this one so much.
Supporters of truth, insight and of data visualizations are not so keen.
Of course, I’m talking about the classic election map. This will look familiar to all. A standard map of the USA – like the one seen in schoolrooms and public offices all over the country – colored in red for Republican and blue for Democrat.
What’s So Bad About It?
“What’s so bad about the good ol’ election map?”
Our friend at Vox Media have created this video to explore the simple fact that it doesn’t serve the purpose of any good data visualization: to accurately represent the date.
And it may have impacted the outcome of the election.
They point to the representations of Massachusetts and Montana on this map as an example of why this visualization is so, so wrong.
To put it simply, Montana is big. The fourth largest state in terms of area covers over 147,000 square miles of mainland USA and is over 630 miles long from east to west. Put a state like this on a map, color it in red, and that red becomes dominant.
However, by other definitions, Montana is small (sorry Montana.) There are only 1.04 million residents in Montana – the 6th lowest in the nation. As a result, Montana has only three Electoral College votes.
Contrast this to The Bay State, Massachusetts, nestled into a corner of the east coast, sandwiched between other tiny, blue, well-populated states.
In a nice bit of symmetry, Massachusetts is ranked sixth lowest in the nation in terms of area, tipping the scales at only 10,565 square miles. However, the state’s population is over six times that of Montana’s, making it more significant in terms of a representation of the entire US populace.
This means Massachusetts gets 12 Electoral College votes to Montana’s three. But on the map? Massachusetts is buried. There is no disrespect here, Montana is a fine state. But in terms of influence and significance in an election, it is less prominent than others.
Check out the Vox Video below:
So, what are media outlets and visualizers trying to achieve with this map? The answer to this depends on how much cynicism you want to inject.
At best, visualizations like these are lazy attempts at giving audiences a visual reference point and an array of data for them to consume.
Thanks to the problems with scale and the fact that the physical size of a state has nothing to do with any of the data parameters, this attempt falls flat.
I can almost understand it. After all, audience members want to be able to quickly identify key states and understand the result in that location. Media outlets are afraid that, by skewing the sizes of the states, the audience will be confused.
However, a more cynical approach can be taken. The big midwestern states tend to vote somewhat differently to the relatively small states in the northeast.
Applying the Principles
Ok, so let’s imagine a world in which we have done away with such visual abominations. What is in their place?
One possibility is the cartogram. A cartogram is a map on which the physical sizes of the states are distorted to represent their population and electoral significance. Here’s one prior to the the 2012 election:
This approach is already being used by the New York Times and delivers a quick and easy reference point without doing away with the idea of a map altogether.
However, it has not yet been widely adopted, perhaps because the resulting map is not instantly recognizable as the United States.
To derive other ways of representing this data, we simply need to apply the principles of DataViz. The visualizations must be genuinely useful, it must facilitate a stronger connection with the data than merely reading it from a table, and it must never, ever distort the facts.
Whether we decide to use a bar chart, cartogram, or some other method of presenting vital polling data, one thing is for sure; the misleading maps of old need to remain, where they are, buried in a shallow grave in 2016.