Visualizing with Clarity - Data Points: Visualization That Means Something - page 557

For example, the United States has a two-party system, with Democrats and
Republicans. Blue is the color of the Democratic party and red is the color of
the Republican party. Therefore, a map, as shown in Figure 5-14, should relect
the party colors. Flip the colors, and the proportions between two groups
would be the same, but because the party colors are so commonplace, it's
probable that readers would misinterpret the results as a Barack Obama win
in the Midwest and Southeast and a Mitt Romney in the West and Northeast.

The charts in Figure 5-15 show movie tril-
ogy ratings from the review aggregation
site Rotten Tomatoes. On the site, a ripe red
tomato is used for movies that earn at least
60 percent positive reviews (fresh), whereas
a splattered green tomato is used for movies
below the 60 percent threshold (rotten). The
graphs match the site's color scheme so that
you can easily see which movies were fresh
and which were rotten. The length of each
bar provides a more exact value.

Context can also afect your choice of geom-
etry. For example, the Bureau of Labor statis-
tics releases monthly estimates for number of
jobs lost and gained. Figure 5-16 represents
jobs lost between February 2008 and February 2010. More jobs were lost than
gained every month during this period. The taller the bar is, the more jobs
that were lost on the corresponding month.

FI G U R E 5-14 Color by expectation

The chart with values in the positive makes sense, but consider the context
the chart is usually presented in. People expect to see bars in the positive
for jobs gained and in the negative for jobs lost. However, the coordinate
system in Figure 5-17 would put jobs gained in the negative. Negative jobs
lost means new jobs.

So instead, it's more intuitive to frame jobs lost as negative values, as shown in
Figure 5-17. It makes more sense to show something lost moving downward,
when that something is looked at negatively. On the other hand, decreased
weight, when the goal is actually to lose weight, might work better on the
positive side of the axis.