by: Matthew Holm
From Research/Penn State, Vol. 16, no. 3 (September, 1995)
side from the occasional impromptu napkin scribble, the
last time I had to draw a map was in second grade, when Mrs.
Heimer gave us brightly colored workbooks that showed just
how it was done. One hour, three afternoons a week, we drew
the various symbols guaranteed to be useful for life: tiny
trees for forests, wavelets covering splotches of blue for
lakes or seas, and my favorite, the bridge. It looked
something like two reversed brackets ][ but Mrs. Heimer
promised it was universally read as "bridge."
I've never needed to draw a bridge since, and I've
never once seen waves gracing the lakes and seas of Rand
McNally. Professional cartographers would seem to have their
own set of rules.
But as Cindy Brewer of Penn State's geography
department tells me, map makers don't necessarily understand
how people read maps, and often draw them in confusing ways.
Her work has been to make maps easy to read and easy to
draw. "I'm trying to be real systematic," she says. "What do
you do with color to have the map make sense?"
Brewer works not with road atlases, but with thematic
maps of death rates, employment rates, or weather data,
for example. Color choice in this type of map is very
important, because the relationships between the colors
determine how easily data trends can be seen.
Brewer understands color: her undergraduate work at the
University of Guelph was not in geography, but art. The art
training made her familiar with color theory, and later, for
her master's thesis at Michigan State University, Brewer
took the values of the Munsell color space which describes
color as having the attributes of hue (whether it's red,
yellow, green, blue, etc.), chroma (the intensity of the
color), and value (light or dark) and made detailed charts
that gave both Munsell specifications and CMYK (Cyan,
Magenta, Yellow, and blacK, the inks that professional
printers use) percentages for the colors, with each hue on a
separate chart, progressing from light to dark and saturated
to dull. These color charts let map makers pick paint chips
that have the colors they want on their maps, and then give
a printer the CMYK percentages to reproduce the Munsell
colors accurately.
With a useful color chart at her fingertips, Brewer
decided to see how else she could simplify and standardize
map making. One problem she noticed was in the color choices
map makers made when showing different percentage values,
they often used colors of entirely different hues. "Rainbow
colors make data trends hard to see," Brewer says. Instead,
she suggests using a single hue but changing its lightness
to indicate the change in the percentage value.
Her research has shown that lightness or darkness is
perceived as a measure of importance: darker colors signify
more, while lighter colors mean less. If the map deals with
things of equal value such as which party controls which
voting districts the data should be represented with
different hues of similar lightness, so as not to imply that
one party is better than another. "But if you're talking
about percentages or values," says Brewer, "use different
color values."
In the map design classes she teaches at Penn State,
her students choose their own projects and supply real data
to make the maps. The class sometimes proves to be
educational in areas outside of cartography, says Brewer.
"Last semester, I had a kid who wanted to show that
motorcycles were safer than cars," she recalls, pulling his
project from a stack of maps in her office. She laughs to
see me start, in my face something of the surprise and
horror her student must have had when he first made the map,
for it displayed quite clearly and graphically how
incorrect his assumptions were.
ne thing Brewer insists on is using real data in her
map-design studies many map researchers simply plug in
faked data in areas that seem appropriate, lending, she
feels, an artificiality to the research. Brewer has always
been concerned with realism and practicality, in both her
research and her teaching: "I guess you could say my work
has a real applied bent to it."
She pulls open one of the wide, flat map drawers that
line one wall of her office: the inside is sprayed with
small, colorful maps. "In the past, I've had my students
make their final projects into postcards," she says,
flipping one over to show the picture caption and tiny stamp
square on its white back side. Also in the drawer are maps
printed on sandwiched acetate sheets. Brewer peels one sheet
back: it is printed entirely in yellow ink. The other sheets
are exclusively of cyan, magenta, and black, but when all
four are stacked together they make a full-color map. "I
really want my students to leave with a piece of
publication-ready design work," she explains.
Brewer and her students do their mapping on Macintosh
computers, making it easy to change colors and create
professional-quality maps. She pulls one up onto the
computer screen to demonstrate her research methods. "One
thing I'm interested in is how the surrounding colors affect
the color you're looking at," she says, adding that her
Ph.D. work at Michigan State was in this area. Brewer points
to a pale blue section of the map and asks me to match it
with the other blues on the map. The task isn't as easy as
it sounds, because surrounding dark browns tend to make the
blues appear lighter while surrounding yellows make them
seem darker and bluer. After some quick double-guessing
based on my own painting experience, I choose the correct
color from a set of blues on the map. Brewer then switches
maps, points to a yellow-green blob, and asks me to match
it. "Match it?" I ask. "You mean there's more than one
green?" She laughs as she clicks the pointer on the shape,
dragging it on top of a similar green the overlapping
makes the color difference immediately obvious.
The colorblind, of course, have even greater
difficulties with color-recognition. Brewer has been playing
with different color schemes to field-test colorblindness
studies other researchers have done. For example, while a
colorblind individual may confuse red and green, they can
differentiate between pink and yellow-green. By making
slight adjustments, Brewer can make map colors more
distinct.
"But we also have to consider whether nameability of
the colors matters," she says. While the readers might tell
the colors apart, map communication may still be unclear if
they can't properly label them for example, Brewer's
description of the "pink" color above was much more
convoluted at first: "that reddish, pinkish, purple kind of
color."
In addition to color perception, Brewer asks her test
subjects specific questions about the maps, to determine how
they interpret the data. She asks about basic matches (what
percentage is shown in a specific area, for example),
whether the levels are higher in one area than another, and
if the data is clustered into regions.
"Epidemiologists are very interested in that," Brewer
says of the clustering they study clustering, and other
patterns of disease spread. Her current project, with Alan
MacEachren of Penn State, is to advise the assembly of an
atlas showing mortality rates throughout the country for the
National Center for Health Statistics, one of the Centers
for Disease Control (CDC).
Brewer is currently testing a variety of mapping
techniques for the CDC project, to determine for certain
which ones will be the most clear and effective to the
common reader. "We want an atlas that even your colorblind
congressman can read," laughs Brewer.
But after the CDC project, Brewer would like to examine
how people who are experts with the data read maps, as
opposed to how lay persons do. Possible subjects might be
epidemiologists, or environmental researchers studying
global change: "people who really want to figure something
out, and know their data," she says.
For both experts and novices, Brewer's research could
make maps more universally understandable. Her
standardization of map conventions could serve as a ][
between those who draw maps and those who read them.
Cindy Brewer, Ph.D., is assistant professor of geography,
325 Walker Building, University Park, PA, 16802; 814-865-5072.
Alan M. MacEachren, Ph.D., is professor of geography,
310 Walker Building, University Park, PA, 16802; 814-865-7491.
The mortality rates project is being funded by the
National Center for Health Statistics.
