Data is the most fundamental piece of cartography. From detailed numeric records such as population or the economy to simple measures like time and place, data represents the very bones of every map. How ever, misunderstanding data within maps pervades our society. By revealing many of the challenges that face cartographers on a daily basis, we can all become smarter and more ambitious map enthusiasts.
A photograph captures the very essence of a moment. The environment, abstract artistic vision, and raw emotion come together to create an experience that is fundamentally human. It is well worth the one thousand words thrust upon it by the vernacular idiom. However, if a picture offers a thousand words, then surely a map tells a whole story. Of course, we are not talking about literal words. A photograph has a language unto itself: color, light, ambience. A map is no different, though the dialect differs. In maps, the common tongue is data. Just like the spoken word, data can tell fantastic stories and uncover new meanings. However, when misunderstood, it may turn into a language of manipulation and ignorance. To become fluent with maps and the data with which they communicate is one of the most fundamental challenges for cartographers and map enthusiasts alike.
To Map or not to Map
All data has a natural visual form and it’s vital to the integrity of the data to discover it. Your information may need to be a map to fully express itself or perhaps a simple scatter plot would suffice? Seeing how much ice has been lost at the North Pole is easy to communicate in a map. While, putting that same data in a pie chart may be accurate but less compelling. Why deprive the reader of a visualization that the data is able to provide? This can be a dilemma for cartographers because the truth is, not all data should be a map. Though data may have coordinates or a location, mapping the data may not be the best way to tell the story.
One habit among designers that leads to poor maps is the overreliance upon political boundaries. When data is constrained to a nation’s borders, you may be working with sizes that don’t necessarily reflect importance. Imagine creating a bar graph where the widths of the bars vary for no reason; this graph would be unnecessarily confusing.
On the other hand, when you’re mapping something natural that flows across the map without barriers, data thrives. Information such as weather or wildlife migration is less sensitive to distortion than country data, for example. Our chart of firearm ownership versus homicide rates in developed countries shows an alternative to bad political boundary maps: If this were a political map the tiny countries of Denmark and Singapore wouldn’t get the prominence it deserves, while Canada and Australia appear deceptively prominent. The map was also only able to display two variables (gun ownership and homicide rate) whereas the chart includes three (the two mentioned plus gun homicide proportion). This may lend a reader to accidentally make the wrong conclusion. If you need to spell these nuances on the map, then perhaps it wasn’t the best choice to visualize the data in the first place.
It’s important to note, though a map may not be the best choice, that doesn’t mean that other visualizations can’t be geographic. Charts such as our Gender Equality infographic and Climate Change by Latitude graph express data dependent on location using alternative means from a map. This gives the data the space it needs to be legible and effective.
Maps tell stories, but the story only works if the “sentences” are coherent.
Exploring vs. Persuading
When starting to look at data, it is important to determine whether that data is trying to persuade the reader or set up a space for exploration to allow conclusions to be made organically. For example, National Geographic produced a map of Virgin Mary sightings. The map was to be explored across the page; there was no agenda to the image other than what the reader personally took away.
Persuasive data, for example, might be used to map deforestation or racial segregation. When working with persuasive data, the map is telling the reader, “There is a problem here,” and the goal is to lead the reader to that conclusion.
Often, maps are both exploratory and persuasive. Maps have a spectrum of data and that data may be able to express itself without the cartographer holding the reader’s hand. Our map of voting booths in Saudi Arabia exemplifies this strategy. The issue with women not having reachable voting booths is not highlighted on the map, but as the reader looks at the ratio of men’s to women’s polling stations they can reach the conclusion themselves.
Gaps in data, while problematic for all maps, are especially harmful to a persuasive map. When data is missing in a purely exploratory map, there are no results to skew, and these vacancies are easy to forgive. With maps that lead to a conclusion, gaps in the data are the enemy. A cartographer must discover what data is missing and how it could skew the data. Also important is letting the reader know what is missing, because nothing is worse than a critic claiming you intentionally hid something to mislead the reader (though they often will anyway!).
Scope of the data
Scope of data is extremely important as well. Data can come in so many forms and so many levels you may not realize that the results are skewed. Scope may be deciding to express an issue at a local level rather than a global level, or averaging temporal data weekly or monthly rather than yearly. One may make this decision to save time and collect only a month’s worth of data rather than a year’s worth. But making this decision might skew the data more than you’d expect.
One example of this is Atlas Lens’ Syria map, which showed how many refugees fled from a given country. In order to use the most recent data, data was taken from the month of July. This gave readers a sense of the most recent trends of Syrian migration into Europe. However, if the data was averaged for multiple months or even years, different conclusions could have been drawn. What have migration trends looked like in total? How have they changed over time? These tell fundamentally different stories, neither right nor wrong, but it shows how a simple decision of parameters can alter how people perceive a data visualization or map.
Dealing with barriers
Sometimes, the way data is displayed can obstruct its true story. For example, a chorpleth map is one that displays shapes color-coded to a certain data value (e.g. a world map displaying average GPD by country).
Simply put, the choropleth is overrated. It is an easy way for beginning cartographers to attach data to place, and at times it gets the job done. On a world scale though, it is very often misleading. In a choropleth map, country size varies so much it is difficult to make the comparisons you hope to accomplish, and boundaries emphasize areas that may not be important (such as Russia or the United States.)
Choropleths can also create unnatural boundaries for the information. Mapping crime or disease on a choropleth would be a mistake because crime and disease don’t recognize political boundaries. This data would be better expressed as a density-sensitive visualization such as a heat map.
Time and Space
Where is often inseparable from when, and cartographers must remember this. Time can be a crucial piece of information, and leaving it often results in an incomplete map. For example, attacks by Boko Haram or ISIS are time sensitive. When did they happen? Have they become more or less frequent? When mapping this data, a cartographer shouldn’t leave the reader wondering these questions.
To avoid leaving these questions unanswered, one may use color to express time. When colors have already been assigned to another variable, it may be necessary to add a timeline. Size is usually not a good way of visualizing time because there are too many connections between the size of something and other values.
In some ways data is an art form, and the choices are those of the creator to make. Though it is important to always ask ourselves why each choice is made and its role in the story. But most importantly, we must be flexible. Visualizing data in a form that does not tell its story right can lead to misunderstandings by readers. Keeping these concepts in mind is the first step to great map making, and vital to great map reading.