The document discusses techniques for effectively formatting bar charts with fire incident data. It provides an example of an over-formatted bar chart and simplifies it through a series of steps like removing unnecessary effects, labels, ink, and whitespace. The simplified chart is easier for readers to understand at a glance. The document emphasizes that charts should be clean and simple to clearly inform readers rather than using all possible formatting options.

1. Fire data isn’t ugly
Presenting fire data effectively series
Episode: Over-formatting
June 2015

2. A makeover of fire department
data to transform it from
unnecessary and confusing to
informational.

3. It’s no secret that a chart can offer quick
insight when a table fails.

4. Microsoft Excel is not just a powerful tool to keep data
and spreadsheets. It’s a friend for creating quick charts.
However, just because you can do something to a chart
doesn’t mean that you should.

5. Exhibit: Bar Chart
“Formatting free-for-all”

6. 0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
#ofIncidents
Number of incidents by time
0 1
2 3
4 5
6 7
8 9
10 11
12 13
14 15
16 17
18 19
20 21
22 23

7. Say it with me now:
Just because you can, doesn’t
mean that you should.

8. The good news is that this chart type is perfectly
fine for the data being represented: call volume
by hour of day.

9. Remove to improve
A popular mantra in the world
of data visualization

10. 0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
#ofIncidents
Number of incidents by time
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
#ofIncidents
Number of incidents by time
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
It’s a small thing that your eye barely notices, but avoid
using effects like shadow, glow, and soft edges
anywhere on a chart.
Here, we’ve removed the shadow (right) to give the
columns some space to breathe.

11. 0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
#ofIncidents
Number of incidents by time
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
#ofIncidents
Number of incidents by time
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
We only have one type of information presented here
(calls) so the columns really don’t need all the colors.

12. 0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
#ofIncidents
Number of incidents by time
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
#ofIncidents
Number of incidents by time
And since we’re only comparing one type of data, we
don’t need the legend at the right.
The lesson here is to get rid of ink that doesn’t help.

13. 0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
#ofIncidents
Number of incidents by time
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
#ofIncidents
Number of incidents by time
Speaking of unnecessary ink, let’s drop the background
texture and color.
Already, the amount of work for your eyes is reduced
and now we can make the chart bigger

14. 0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
#ofIncidents
Number of incidents by time
We don’t need the “beveled” effect on the slides.
Flattening them out gives our eye a continuous line that
doesn’t have to compete with the highlighted top. We’re
looking pretty clean at this point but we can keep going.
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
#ofIncidents
Number of incidents by time

15. The chart title clearly provides what we’re measuring,
number of incidents, so we can drop the left legend. Try
to avoid bold text (title) as it can get blurry, especially
when printed.
I moved the title to the left out of personal preference.
Western writing starts at the left, a habit for our eyes.
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
#ofIncidents
Number of incidents by time
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Number of incidents by time

16. Our maximum calls in any given hour is 244 but the
chart extended to 300. Drop the unnecessary space by
changing the y-axis maximum to 250.
It’s personal preference but I lighten the major gridlines
to a grey, making the columns the stars they are.
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Number of incidents by time
0
50
100
150
200
250
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Number of incidents by time

17. “Truncate” an axis
It’s ok to change the maximum but
never change the minimum from
“0” without doing some extra work.

18. There are times when you need to emphasize small changes
between values. One way to accomplish that is to “truncate”
the Y-axis, which changes the minimum value.
If you truncate, make two charts: one with the Y-axis at “0”
and one with a truncated axis, with a note referencing the
full chart.
0
20
40
60
80
100
120
140
0 1 2 3
100
105
110
115
120
125
130
135
140
0 1 2 3

19. Most people won’t notice if you truncated an axis but the
ones who do will appreciate both charts.
Consider checking out the book How to Lie with Statistics
by Darrell Huff for more.
0
20
40
60
80
100
120
140
0 1 2 3
100
105
110
115
120
125
130
135
140
0 1 2 3

20. “Narrow axes can make small and inconsequential changes seem
big, but-symmetrically- zero-axes can make big and real changes
seem small. What matters isn’t some iron rule like ‘Always have a
zero-based axis!’, it’s your prior commitment to being honest with
the data.”
-Kieren Healy
Duke sociology professor

21. Help your reader
Make sure that you have clear
labels. Don’t assume your readers
will intuitively know what it is they
are looking at.

22. A layman may not realize the hours are in 24-hour
clock, starting at midnight. We can add a legend to the
bottom, specifically stating the measurement.
Working memory generally taps out at 3 items and the
legend simply helps that memory keep rolling.
0
50
100
150
200
250
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Number of incidents by time
0
50
100
150
200
250
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour of Day, 24-Hour Clock
Number of incidents by time

23. 0
50
100
150
200
250
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour of Day, 24-hour clock
Number of incidents by hour of day
We could stop here, but if you have a direct message
consider coming out and saying it right in the chart title
and adding a bit of emphasis.

24. 0
50
100
150
200
250
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour of day, 24-hour clock
Based on our call volume per hour, we might consider moving scheduled shift
changes to 6am, rather than 7am. Call volumes begin increasing immediately at
7am. This would allow more time for the current shift to finish duties, rather
than being constantly on calls up to shift change.

25. 0
50
100
150
200
250
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour of day, 24-hour clock
We can control where a reader’s eyes focus by using color.
In this case, change the color to a medium grey and put a
pop of color on one column.

26. 0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
#ofIncidents
Number of incidents by time
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Before
After
0
50
100
150
200
250
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour of Day, 24-hour clock
Number of incidents by time

27. Over-formatting hurts
Charts are meant to inform, and they
do this quite nicely when they are
clean and simple.

28. Hello! I’m Sara Wood and I love converting fire service members into
NFIRS operatives. I’m currently the State NFIRS program manager for
Kansas and enjoy providing classes and help to bring fire departments
into the era of data driven decisions. If you need help creating a
presentation or analyzing your data, I’d love to hear from you!