Bad Maps that help the spread of the Coronavirus: How the Media misuses Choropleth Maps

Maps are can be powerful tools to be used as a vehicle for the transmission of knowledge. It is said that a picture is worth one thousand words. A well designed map could be even worth many more thousands of words. However, the maps that I have seen used by the media in displaying coronavirus data break important rules of cartographic construction so much so, that they send the wrong messages to anyone who is not schooled in the art and science of map making.

I was fortunate enough to study Cartography (map making) under the tutelage of Dr. Borden Dent over thirty years ago. The lessons he taught his students in map interpretation still stick with me today, and I strive to incorporate a little bit of this knowledge into the various Geography courses that I teach. His course was the first one I took at any academic level where the textbook used in the class was written by the instructor. According to Dr. Dent, the “needs of the map user are very important in the transfer of knowledge between map author and user.” (1) If the goal of the coronavirus maps used in the media are to educate us to change our habits to mitigate the effects of the pandemic, then I think they are failing to do so. From the maps that I see used on the nightly news and in many print publications, it is clear that the people who made those maps either had no formal training in cartographic methodology or they are deliberately using those maps to mislead people. I am not comfortable with either of those possibilities. Therefore, I feel compelled to take you on a short geographical journey into the world of Choropleth Mapping.

In his book, “Principles of Thematic Map Design”, Dr. Dent defines a thematic map as either a special-purpose, single-topic, or a statistical map. There are two kinds of thematic maps; qualitative or quantitative. The main thing that a qualitative map does it to show the spatial distribution of some kind of phenomenon that is not statistical in nature. A good example would be to show the different types of ecological regions on a map. It is either classified as forest, steppe, tundra, or desert. There is no quantity associated with a qualitative map. It is either this, or that. Below is an example.

A qualitative Thematic Map of Bio-regions

The other type of thematic map is quantitative, which shows the spatial aspects of numerical data. Data can be either discrete (individual phenomenon), or continuous (phenomenon occurs between observations). A map of temperature or elevation are examples of continuous data. There is always a value in between observations. Take two cities close to each other which report different current temperatures. One can assume that there is a continuous transition of temperature between the cities connecting the two different values. Population, on the other hand is discrete. Our subdivision is at the eastern boundary of the city. To the east of us there is some open land. There may be no occurrences of people between spatial observations. Maps of Covid-19 are used to show how many cases and where they occur, so therefore they show discrete quantitative data. They use nominal, ordinal, interval, and ratio measurement scales.

The type of map symbols or the type of map the cartographer uses depends on the type of data and the scales used. Data can be either raw data (an absolute number), or derived data (a percentage). Take a look at the following map and study it for a few minutes. What message does it convey to you?

The above map is an example of a choropleth map. The etymology of the name comes from the Greek “choros” meaning place, and “pleth” meaning value. These types of maps are widely used (and misused) by government agencies such as the census bureau, CDC, political parties, etc. because government data is compiled using political boundaries, which serve as enumeration districts for mapping.

There are just a few rules that need to be applied. According to Dent, “the choropleth technique should be selected ONLY when the form of data is appropriate.”(2) It is customary to use derived data (ratios) and not raw data because enumeration districts vary in size. Now, did you surmise that the above map suggested that the coronavirus pandemic was mostly worse for the Northeast states, Illinois, Florida, Texas and California? Did it seem like the problem was not so bad in North Dakota, Idaho and Montana? Why would that be? The reason is the map maker used raw numbers of reported cases instead of using ratios. The populations of our 50 states widely vary. Of course the most populous states will have the most cases and the least populated states will have less!

What conclusions could a map reader take from this map, if they had never been schooled in the art and science of map making? They might wrongly conclude that this is an urban problem, a “blue-state” problem. This has already been talked about for over a month now. Many of the rural states opened up economies early and did not take the pandemic seriously. Now, states like Arizona, Arkansas, and South Dakota are experiencing steep upward curves of Covid-19 infections. States like NY and NJ which was where on the of original epicenters of spiking infections, have now gotten a handle on it. Along with CT, and RI, they not only have flattened the curve, but have seen downward trends in infection rates. But since most have high populations, the map of raw numbers of cases gives the impression that it is still a center of mayhem.

New York State on the downward trend

How then, should we display the data? If comparing states with vastly different population numbers is analogous to comparing apples and oranges, then what should we do? We need to use derived data (rates or ratios). The most common way is to divide each state’s (or whatever geographical enumeration unit you use) population by a number, so that you can compare low population locations with high population locations. If the map you are analyzing shows only raw numbers…..then you, as the map reader should question the map maker’s agenda.

Now take a look at the following map, which shows areas affected by Covid-19 PER CAPITA. I bet you will get a different message….

Covid-19 cases per capita

The darker the color on the map, the higher percentage of population that has contracted the virus. The enumeration districts are now counties and not state borders, which helps the reader see the distribution better. This map still shows high per capita infection rates in the Northeast USA, but now take a look at the Dakotas, Wyoming, the Snake River section of Idaho and some of the Native American lands near the Four Corners region and NW Nevada. These are hard hit places which do NOT show up in the first map. It shows that Covid-19 is neither a blue-state phenomenon, or an urban one.

Many of us have heard the excuse from the president that higher testing means more cases and that we should stop testing so much. As the virus spreads, even the per capita map will begin to fill in with high numbers, since once a person has tested positive they will always be counted in the per capita statistics, even though they may have recovered from the virus. The numbers will continually go up unless we add lots more population either by high birthrates or by immigration. So, what kind of data could tell us if the virus is spreading faster or not? One map that I haven’t seen, but would like to see, is a map of percentage change of positive tests to the ratio of all tests given.

Dr. Dent passed away several years ago. I thank him for the knowledge he imparted to me, both in cartography and in life. I’m also glad he isn’t viewing the plethora of bad public health maps that seem to be everywhere. Although throughout history, some bad maps have misled people for nefarious reasons, most misrepresentation today may be the result of decreased emphasis in Geography in our education system.

People are naturally cautious when they consume news, as they should be. We tend to be skeptical of the written and spoken word, as many news outlets have underlying political agendas; some right leaning and some left leaning. However, when propaganda comes to us in the form of maps or graphics, we are more prone to have the wool pulled over our eyes. The only way to avoid this is to become better map analyzers and map interpreters. Just as we consume news and the written word from many different sources, we need to look at many different maps to get the whole picture. Maps can give us new perspectives on issues and can change the way we perceive and respond to life’s challenges. My hope that you, my dear readers, will not look at maps as being untruthful, but look at them analytically. Don’t be turned off to maps just because there are some bad ones out there. A good map should get you to ask deeper questions. The more you understand about how they are made, you can use a healthy skepticism while you analyze them and learn more about the world that we live in.

We only discussed one type of map today. I could write about map scale, types of projections, cartographic generalization, design elements, etc., but that would take a whole book instead of a blog. If you want the “whole enchilada”, you can read Dr. Dent’s book, “Principles of Thematic Map Design” (1,2), or the newer edition “Cartography-Thematic Map design.” There are several other books on how to interpret maps. Another one I recommend is Mark Monmonier’s “How to Lie With Maps” which shows historical examples of how cartographic ignorance shaped the world. Judith Tyner also wrote a good book on map design. Whichever resource you choose, go educate yourself more about what goes into making maps and you will be on your way to graphic and geographic literacy.

UPDATE: September 2021…..For better maps on Corona Virus stats, go to ourworldindata.org. They have proper choropleth maps which use derived data and not raw data.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: