The maps are out of control

 •  Filed under Data science and visualization

Even if you love geography, you probably do not click on anything that promises to explain "everything"...

As Danny DeBelius says: "nothing is certain in this life but death, taxes and requests for geographic data to be represented on a map."

Folks in the chart-producing industry seem to realize that what we want to communicate is often the range of some variable because readers care about disparities. Sometimes they may also care about geographic dispersion, but often they will not. Accurately visualizing the exact distance from the equator is typically not relevant for the story. So Danny DeBelius let his readers vote whether hexagon tile grids, square tile grids, or geographic choropleth map were most effective:

That's a great poll to run. I hope that the most difficult option will win - that way, the rate of growth of the total number of maps produced may decelerate, and the share of maps that are useful will increase. (Semi-serious.)

Also, imposing a square-based structure on a map is helpful when somebody insists that you must make a cartogram. In the example below, the size of the square reflects population (the colors seem to be chosen only to distinguish between countries).


But as one commenter wrote a while back, cartograms are "asking us to compare the relative size of odd shaped areas (countries scaled by data) to a memorized area (the non-scaled world map). "

In most cases, we still use colors to highlight differences between countries and regions. But even then it's essential to keep the map readable, especially if the whole world is shown. This example works fairly well:

Still, a map should not be the default choice, just because the data comes with geographic identifiers. Sometimes, more than half of the country will look empty:


Or you won't have any data for a large share of the land mass, if your survey is not global:

Source: BBC

Other than the "empty spaces", the map has a vague scale: how much better is Mexico than Brazil? Is the difference meaningful? And are the categories relative, or would every country have been dark green had students done well on the tests everywhere?

Finally, what should be make of this infographic?


It seems to want to tell us an alarming story about the national obsesity trend. So why are we looking at differences between states?

Can we do better?

Given the poor readability and aesthetics of many maps, these submissions that were made as a joke would actually be an improvement of many existing visualizations...

People will surely find ways to make maps more fun and, in some cases, informative. But in spite of the oversupply of maps on news sites recently, virtually any map would be better than what Le Monde just did.