How-to-Tell science Stories with Maps

Reported Featureshow to tell Science Stories with MapsAugust, Greg Miller .

This map, part of the Audubon ' s Birds and Climate, depicts predicted changes in tree swallows ' summer (yellow) a nd winter (blue) ranges.

Maps is amazing for their ability to show us something we can ' t see directly, from the path of the Curiosity rover on Mar s, to the tangle of underground fracking wells in North Dakota, to clusters of unvaccinated schoolchildren in California. For journalists, the maps can be both a powerful data-visualization tool and a reporting tool.

Maps is some of the most information-dense ways of communicating data, says Len De Groot, director of data Visualiza tion at the  Los Angeles times . People understand maps intuitively because they use them in their everyday lives, De Groot says. "You can do a IoT in a map because people already understand the fundamentals-unlike, say, a scatterplot."

Maps can also reveal relationships and stories that aren ' t otherwise apparent. In the mid-2000s, De Groot is part of a team at the  South Florida sun-sentinel  that mapped FEMA Disbur Sements after several hurricanes, including Hurricane Frances, which struck in 2004. "We didn ' t start with any agenda, we were just doing the standard where ' s-the-money-going thing," he says. "To our surprise there is one zip code in Miami where we saw there is a spike in payouts in areas where we knew there WA s very little damage. " That led To a broader investigation by the paper, which revealed widespread fraud and got the paper nominated fo R a Pulitzer Prize and ultimately led to policy changes at FEMA.

Fortunately for cartography-curious journalists, making maps are easier than ever, thanks to a proliferation of (mostly) FR EE and Open-source software. Anyone with a computer can put together a nice-looking map in less than an hour. Making really good maps isn ' t quite that's easy, however. It requires more thought and a willingness to learn some technical skills. Here is some considerations to keep in mind, and some tips on how to get started.

Whether to Map

Just because you  could  make a map from a given datasets doesn ' t necessarily mean you should. The key question is whether are the most important thing about the stories you ' re trying to tell, says Virginia Maso N, a senior graphics editor at  National Geographic . For example, one upcoming stories involves a Department of agriculture program that culls wolves to protect farm animal S. The graphics team had data on how many animals has been killed in each state, and their first thought is to make a MA p, shaded to indicate the number of animals killed. But they soon thought better of it. "Really the significance isn ' t  where  the states is, it ' s  what  the states is," Mason says. So they decided to instead create A chart that makes it easy-to-see at a glance which states has killed the MoS T Wolves.

If the same data were displayed on a map, readers would has to scan the map and make comparisons between the colors or Shading of different states. It ' s something our brains aren ' t especially good at, says Peter Aldhous, a self-taught mapmaker and science reporter FOR&N Bsp BuzzFeed News . There ' s been a lot of perceptual science in what visual features our brains is best at discriminating, aldhous s Ays. It's found, for example, so we're much better at comparing the length of lines than we're at comparing area, color, or Saturation. So, if comparing numbers are the point, a bar chart is a solid choice. (For a good introduction to using perceptual. Better infographics, check out Alberto Cairo ' s book, the functional Art ).

Mason and her colleagues at National Geographic do make a map, or rather a series of them, for another Conservat Ion stories, a feature on African lions. The maps show the historical range of lions, which covers large swaths of Africa, and the scattered pockets where they sti ll exist today. In the case, she says, place matters-the Lions now survive almost entirely in wildlife parks and other protected areas. Also, most readers aren ' t as familiar with African geography as they is with the states. "If You do a chart of lion populations by wildlife reserve, people aren ' t going to know where those reserves is," Mason Says.

This map, published in National geographic in, shows the historical range of lions in Africa and the scatter Ed pockets where they still exist today.

What's Your point?

Once you ' ve decided that a map are in fact "what are you want, the next question are what point you want your maps to get across. Then, spend some time thinking the "What to Design" the "map so" every part of it emphasizes this point, and anything th At doesn ' t goes away, says Anthony Robinson, a geography professor who teaches cartography at Penn state. (Robinson also developed a mapping moocthat ' s been taken by more than 100,000 people-it ' s a great introduction, but the NE XT installment hasn ' t been scheduled yet).

"One common thing I see was satellite or aerial imagery on something like election maps," Robinson says. If the point of the map is a who voted for whom, the land cover is irrelevant. A neutral white or gray background is usually preferable.

Too much clutter is another pitfall, Robinson says. Mapping software puts every little road, stream, and populated place at your disposal. Resist the urge to include them all, even if you had to rebel against the default settings of the Software or remove feat Ures by hand using the Adobe Illustrator or Asimilar program.

Can see a nice example of a strong visual hierarchy on the map below, from the New Yorktimes.

This New York times map documenting the movement of refugees mutes details of oceans and landmasses to minimize C Lutter.

The red lines show the movement of refugees around the world. The white oceans and gray landmasses is just visible enough to provide a reference. "We used a satellite image, but we are took out of all the color so the Red Arrows really popped," says Tim Wallace, a cartograph Er and graphics editor at thetimes. This map also illustrates how making a maps can reveal new stories-in contrast to recent media attention to refugees ' Despe Rate attempts to cross the Mediterranean and Andaman Seas, the map shows a surprising amount of displacement within Africa .

Pick the right Projection

in 2003  the Economist  published a maps to accompany an article on the threat of missile attacks from North Korea. They used a standard mercator projection and concentric circles to indicate the regions potentially in harm ' s WA Y. wrong! The Mercator projection grossly distorts distances near the poles, so the real reach of any missiles from Pyongyang would Be far greater. The magazine published a correction, along with a revised-and much scarier-map a short time later (see both Versionsh ere).

Picking the wrong projection is a rookie mistake. The problem is simple:the Earth was round, and maps are flat. Projections is the mathematical formulas that force a round planet onto a flat map. Each of the projection distorts some combination of area, shape, direction, and distance. Each have its strengths and weaknesses. Mercator, devised for 16th-century nautical charts, was very good at preserving direction.  The azimuthal equidistant projection, which preserves distances relative to a central point, might has been a good choice For the North Korea map. If preserving area was important, say for mapping the impacts of climate in bird habitats, the Albers equal-area con IC projection is a good choice.

Normalize Your Choropleth

When I first met Anthony Robinson, at a mapping conference, he is wearing a little pin that said: "Normalize y Our Choropleth. " I had no idea about it meant, but it turns out to be extremely important advice.

Choropleth maps is everywhere. They use colors or shading to map a variable-anything from election results to crop yields to cancer rates. But if you're not careful, it's easy-to-end up with a misleading or meaningless map. Take cancer. If you map the number of cancer cases in different areas across the U.S., you'll likely see big dark splotches along the Eastern Seaboard and Southern California, with a few small blobs across the Midwest. Because that ' s where the people is. You might as well has made a population map. What's far more interesting is where the cancer rate-the number of cases per 100,000 inhabitants-is higher or lower than n Ormal. Normalizing (by population, in this case) allows readers to make meaningful comparisons. In other cases, you might want to normalize by Area-for example, a map of air-pollution sources might is normalized to sho W The number of factories or automobiles per square mile.

Color Matters

Nothing makes professional cartographers roll their eyes in dismay like a rainbow color scheme, the kind your often see On weather maps. "It's like this trend of putting bacon on all kinds of the food," says Robinson. "It ' s just gratuitous." A rainbow palette may look pretty at first glance Butit ' s rarely a good choice for maps. It's especially bad to mapping a continuous variable like temperature. If Red is warmer than blue, what is Violet (a mix of red and blue) colder than blue? And why are it that a two-degree difference in temperature can correspond to either a subtle change from light green to Dar K Green or a dramatic jump from green to yellow? A more intuitive color scheme would use darker shades of a single color to represent higher values.

Robinson ' s Penn State colleague Cindy Brewer have spent much of her career applying perceptual in color to Mapmaking. Her Website, color Brewer, was a great resource for beginners. It forces think about the data "re trying to map. For Divergent data-areas of Above-and below-average rainfall, for Example-color Brewer might suggest a Color scheme that Transitions from brown to green, with neutral white in the middle. For mapping distinct categories, such as of the dominant invasive species in different areas, the site would suggest a Combina tion of unrelated colors that is easy-to-tell apart. It also limits categories-any more than and readers would has a hard time telling the colors apart. Color Brewer also lets you click boxes to limit yourself to color schemes the can is read by color-blind people, or that Look good on paper as well as on screen.

For a recent anniversary stories on Hurricane Katrina, BuzzFeed News Science reporter Peter Aldhous took Storm Surg e data from a NOAA computer model (left) and tried to the "make it look less like the acid trip in a paint factory." His to subdued but more intuitively readable version was on the right.

Getting Started

Just a few years ago, getting started with digital maps meant diving to the deep end with the kind of complicated geo Graphic-information-system (GIS) software typically used by Natural-resource managers, city planners, and other experts. It ' s much easier these days. One good option Iscartodb, an online mapmaking tool, a allows you to upload datasets and make Basic interactive web maps Using a Point-and-click interface. Sophisticated maps, or to customize your maps so they look exactly as you want, you'll need to learn a bit of Coding. You won ' t become a expert overnight, but can quickly pick up enough to being useful, especially if you take advantage of Cartodb ' S online courses, which were designed with journalists in mind. Pricing varies according to what much data you ' re allowed to upload-a free version allows up to 50MB, which are plenty to pl Ay around with.

For some of my very first efforts at making maps, I used a even easier, Option: arcgis online, from Esri, the MICR Osoft of GIS mapping software. Journalists can request free access through Esri ' Smedia relations team. The Point-and-click interface includes easy access to census data and other public databases, and includes several Templat Es for incorporating photos and text into ' story maps, ' Like this one i made to go with a stories on a wildlife di E-off in Florida. The free version offers less customization and less data analysis than CARTODB, but it's a good option for the code averse .  investigative reporters and Editors recently posted a helpful gallery of video introductions to several m Apping and data visualization tools.

Cartodb, Esri, and other makers of mapping software often offer workshops at journalism conferences. As does the journalists who use them. Aldhous and De Groot has taught a mapping workshop at the Knight Digital Media Center at the UC Berkeley graduate School of journalism. Aldhous have also taught mapping to journalists on the annual meeting of IRE, an organization he credits with fostering he Own interest in mapping and on data journalism more broadly. "They talk a lot on a data frame of mind, where you ' re not just thinking on who can I interview for a story that you ' re Thinking about what can I use data to help me find a stories or bring context to a story, "he says. "The maps is really part of the."

Greg Miller

Guest contributor Greg Miller is a science journalist based in Portland, Oregon. Previously, he is a senior writer at Wired, where he co-founded the cartography blog Map Lab. Follow him on Twitter @dosmonos.

Image Credits:

Tree Swallow map:national Audubon society/stamen Design.

Lion Map:virginia W. Mason/National geographic.

Refugee Map:sergio Peçanha and Tim Wallace/the New York times.

xkcd map PROJECTIONS:XKCD (CC by-nc 2.5).

NOAA and BuzzFeed maps:peter aldhous/ BuzzFeed News and NOAA.

Photo of Greg Miller:courtesy of Greg Miller.

How-to-Tell science Stories with Maps