Dos and donts of figures and tables

1
Dos and donts of figures and tables

• FISH 323

2
Useful References

• Preparing Scientific Illustrationsby Mary Helen
  Briscoe. (Second edition, Springer-Verlag, 1995).
  
• Tufte, E.R. 1983. The visual display of
  quantitative information. Graphics press,
  Cheshire, Conn. 197 p.
• Freese, EW and Miller, RB. Designing effective
  graphics. North American Actuarial Journal 2(2)
  53-76.

3
The Bible
4
Graphical displays should

• Show the data
• Induce viewer to think about substance
• Avoid distorting the datas message
• Present many numbers in a small space
• Make large data sets coherent
• Encourage the eye to compare different pieces of
  data

5
Anscombes quartet of data
6
Each data set has many common properties

• 11 data points
• Average X9.0 Average Y7.5
• Regression line is Y3.5X
• Sum of squares 27.5
• Correlation coefficient 0.82
• All traditional descriptive statistics are the
  same

7
(No Transcript)
8
Graphics show data much more effectively than
tables

• Graphs should show two variables at the same time
• The scatter plot one against another
• The data map spatial data
• Time series
• Frequency distributions

9
Cigarette consumption 1930 vs lung cancer death
rate 1950
10
The scatter plot

• You see the relationship between two variables
• You see observations that appear to be outliers
• Labeling data points, or at least outliers
  contains extra information

11
The data map

• Lets you show observations of intensity etc
  against geographic location
• Particularly useful in ecology where space is
  very important

12
(No Transcript)
13
(No Transcript)
14
The time series

• Try to show more than one thing if possible

15
(No Transcript)
16

• Shows observations and predictions vs time
• Shows deviations between observed and predicted

17
Graphical excellence

• Well designed presentation of interesting data
  substance, statistics and design
• Complex ideas communicated with clarity,
  precision and efficiency
• The greatest number of ideas in the shortest time
  in the smallest space
• Almost always multivariate
• Shows the truth about data

18
Graphical integrity

• Avoid distortion and lying with the graphs
• Area graphs are very dangerous
• Selecting portions of data series

19
Coho salmon return to Puget Sound
20
Maximize data ink
21
Common non-data ink

• Excessive tic marks and labels
• Grid lines
• Unnecessary axis

22
Age distribution of the population of France
23
Principles of data ink

• Above all show the data
• Maximize data-ink ratio
• Erase non-data-ink
• Erase redundant data-ink
• Revise and edit

24
Lets improve the graph
25
(No Transcript)
26
Does this contain as much?
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)