we expect time to be on the x-axis. And most of the data that
EHS people show is time-series data. So the point is to seize on
that presumption whenever possible by turning the page sideways, making it nice and big, and having time on the x-axis.
TS: There are several different kinds of graphs. How do you
know which is the best one to use?
BE: That’s always a challenge, but my advice is, if you don’t
know what to do and the data is over time, the best place to
start is with a bar chart. People often ask about pie charts, and
pie charts are okay for some things. If we have a graph showing “injuries by type in year 2009,” a pie chart showing that
breakdown is fine.
But let’s say we now want to compare 2009 with 2010. Often
we’ll show two pie charts next to each other, but this doesn’t
connote the overall number of injuries. For example, if we had
a hundred injuries in 2009 and 200 injuries in 2010, then one
pie chart should be twice the area of the other. But when you
make a circle twice the area of another circle, it doesn’t look
like it’s twice. If it’s in a bar chart, though, it does look like it’s
twice.
A pie chart to show a snapshot in time is okay, but you
shouldn’t use it to connote orders of magnitude.
TS: How long does it take to make an effective graph?
BE: Most of them probably take me an hour or two. I sit down
with a piece of paper and sketch out what I want to do first.
After I kind of get the vision, it doesn’t take that long.
But that’s certainly one of the criticisms—“I don’t have time
to do all this.” My response is that you don’t need to do this for
all of your graphics. You just need to do it for the one or two
that you really want to tell your story.
Ed Rutkowski is managing editor of The Synergist . He can be reached at (703)
846-0734 or erutkowski@aiha.org.
Don’t Be a DRIP: Quick Tips for Communicating
with Data
Are your graphics DRIP (data-rich, information-poor)? The
following suggestions will help improve your ability to communicate with data:
1. When you’re unsure which kind of graph to use, start with
a bar chart.
2. Don’t rely on the automatic formatting provided by
spreadsheet and presentation software.
3. Make graphs horizontal—50 percent wider than tall is a
good rule of thumb.
4. Make the data and the x-y axes the most prominent features of your graph.
5. Encourage viewers to compare different data.
6. Representations of numbers should be directly proportional to their numerical quantities.
7. Use clear, detailed and thorough labeling.
8. Sketch your ideas on paper before you begin creating a
graph.
Figure 2. A slideshow in the digital edition of the December
Synergist will explain step-by-step how Emery transformed a
graph charting the incidence of nuisance fire alarms at UT-Houston (top) into a much more coherent display (bottom).
Resources on Information Display
By John Tukey:
• Exploratory Data Analysis. Addison- Wesley: Reading,
Mass. (1977).
• “Summarization: smoothing; supplemented views” in
Interpreting Multivariate Data (Vic Barnett, ed.) Chichester,
U.K.: Wiley (1982).
By Edward Tufte:
• The Visual Display of Quantitative Information (2nd ed.)
Cheshire, Conn.: Graphics Press (2001).
• Envisioning Information. Cheshire, Conn.: Graphics Press
(1990).
• Visual Explanations: Images and Quantities, Evidence and
Narrative. Cheshire, Conn.: Graphics Press (1997).
By Robin Williams:
• The Non-Designers Design Book: Design and Typographic
Principles for the Visual Novice. Berkley, Calif.: Peachpit
Press (1994).