Five questions to ask when you see a coronavirus map
Maps help us understand what’s going on as the coronavirus pandemic spreads across the globe. Here’s what you need to know in order to read public health maps with an informed eye.
1. Where did the data come from?
When we see statistics plotted in symbols on a map, it’s easy to assume that they represent a perfectly accurate portrayal of reality. But even the best coronavirus maps are based on data sets compiled from a variety of reporting sources. Each case has to be counted, tabulated, and documented by doctors, public health officials, and institutions. Different political jurisdictions have different capacities for recording cases, so a low number might not mean that there really are few cases—it could mean that few tests are available or that a country or state has ineffective reporting systems. Sometimes other sources, like newspaper reports or obituaries, are useful when public reports are unavailable or inaccurate.
A good map cites its data sources and provides links back to the original sources that were used to create the map. This Johns Hopkins map includes a link to a blog post where the authors detail their data sources. For more on data sources, check out the COVID Tracking Project, a volunteer-led effort to combine and standardize case reports in the US. For each state’s reporting, they assign a data “grade” based on how reliable the reported numbers are.
2. When was the map last updated?
A public health crisis like a pandemic is fast-moving, which means that maps and information published just a few days in the past can be misleading—or even dangerous. Many online maps are meant to be constantly refreshed with new data, but, if they don’t have a timestamp, they may trick us into thinking that they have current information when that’s not actually the case. And as screenshots and images of maps circulate on social media and in emails, they may continue to pop up days or weeks after they went out of date. For maps like these, it’s a good idea to treat them like perishable groceries: if you can’t figure out the expiration date, it’s better to toss the map out and look for one that you know has a more recent timestamp.
3. What geographic units are being used to tally disease cases?
Any geographic phenomenon which takes place at a specific location, like individuals who have contracted coronavirus, is called a point feature. In order to tally up those point features into total counts, we have to choose what geographic unit will serve as our container. Should we count by city, metropolitan area, state, or country? What do we do about the fact that different countries have different types of administrative units? How can we standardize across the different geographic units that are being used to report cases?
It’s also important to remember that different geographic units have different total population numbers. In the Boston Globe map above, Middlesex County looks particularly striking, with 199 cases. But Middlesex is also the most populous county in the state, with 1.5 million people, giving a case rate of around 1 for every 7,500 residents. By contrast, Suffolk County, which includes the City of Boston, has fewer total cases, but also a lower population than Middlesex, so its per capita rate is actually higher: around 1 for every 6,700 residents. Dividing a statistic by population count is called normalization. But even here, the mapmaker needs to make choices: maybe it makes more sense to report the number of cases divided by hospital beds, or some other denominator.
There’s no perfect way to deal with the problem of aggregating and normalizing, but you should always pay attention to what choices are being made on the map—and how they might lead to different conclusions.
4. What kind of symbolization is being used? Is there a legend?
Two visual methods are most commonly used to show counts and magnitudes on a map. A choropleth map uses colors to fill in an entire geographic area according to its data value. A proportional symbol map uses a symbol like a circle, usually centered on top of the area being measured, whose size varies according to the data value.
One disadvantage to a choropleth map is that places with large areas look important, even if they’re not representative of the overall state of affairs. For example, large states like Wyoming, Montana, and Alaska take up a lot of space and ink on a choropleth map of the US, even though they have small populations. For something like a disease epidemic that relates to people, not land area, a choropleth map can be a misleading design choice.
A proportional symbol map is better—both the Johns Hopkins map and the New York Times map use circular symbols sized according to the number of cases—but it’s still important to have a legend so we can understand what magnitudes are being shown. Look at the four maps of Massachusetts above. They all show exactly the same data. But you might draw different conclusions about the severity of the pandemic from each of them!
5. What is the context? What headline and other information accompany the map?
Since news, rumors, and especially visual information can travel so rapidly across social media platforms, it’s important to read every map in its context, asking questions like: Who published this? How are they describing the map? What actions are the authors asking me to take?
The map above was taken out of context from a study looking at travel networks as a way of predicting where the pandemic might spread. But look at the words that The Sun used to frame the map: “horrifying” and “unstoppable,” suggesting that “the incredible graphic reveals” a shocking truth about the disease’s “deadly tentacles.” Remember: maps and graphics never reveal anything all by themselves. They need interpretation and context in order to provide meaning.
More about critically interpreting maps
At the Leventhal Map & Education Center at the Boston Public Library, we’re about to launch a new exhibition called Bending Lines: Maps and Data From Distortion to Deception, which is all about how maps have been used to shape the truth, both historically and in the current day. While our current schedule for opening the exhibition is pending a decision about the library’s reopening, come visit us this spring/summer for more.
