Principles of Graphical Excellence

Best Paper: ALAIR April 5–6, 2001

AIR: June 2-5, 2002, Toronto

Focus-IR, February 21, 2003

Anna T. Waggener, Ph.D.Institutional Assessment

United States Army War College

The Visual Display of Quantitative Information

Leading authority: Edward R. Tufte

History of Graphical Development

• First geographic maps were drawn on clay tablets.

• 17th Century: combined map skills and statistical skills to construct maps.

• Trade winds and monsoons on a world map.• Chart patterns of disease.• Later sophistication showed distribution of

1.3 million galaxies.

“Graphical excellence consists of the efficient

communication of complex quantitative ideas.”

• Organizing Numerical Data:• The Ordered Array and Stem-leaf Display

• Tabulating and Graphing Numerical Data:• Frequency Distributions: Tables, Histograms, Polygons

• Cumulative Distributions: Tables, the Ogive

Presentation Topics

• Tabulating and Graphing Univariate Categorical Data:

• The Summary Table• Bar and Pie Charts, the Pareto Diagram

•Tabulating and Graphing Bivariate Categorical Data:

• Contingency Tables• Side by Side Bar charts

• Graphical Excellence and Common Errors in Presenting Data

Presentation Topics (continued)

“At their best, graphics are instruments for reasoning

about quantitative information.”

2 144677

3 028

4 1

Organizing Numerical Data

Numerical Data

Ordered Array

Stem and LeafDisplay

Frequency DistributionsCumulative Distributions

Histograms

Polygons

Ogive

Tables

41, 24, 32, 26, 27, 27, 30, 24, 38, 21

21, 24, 24, 26, 27, 27, 30, 32, 38, 41

2 1 4 4 6 7 7

Organizing Numerical Data:

•Data in RawRaw form (as collected):24, 26, 24, 21, 27, 27, 30, 41, 32, 38

•Date OrderedOrdered from Smallest to LargestSmallest to Largest: 21, 24, 24, 26, 27, 27, 30, 32, 38, 41

•Stem and Leaf display:3 0 2 8

4 1

“Design is choice.”

O g ive

0

20

40

60

80

100

120

10 20 30 40 50 60

0

1

2

3

4

5

6

7

10 20 30 40 50 60

2 144677

3 028

4 1

Tabulating and Graphing Numerical Data

Numerical Data

Ordered Array

Stem and LeafDisplay

Histograms Ogive

Tables

41, 24, 32, 26, 27, 27, 30, 24, 38, 21

21, 24, 24, 26, 27, 27, 30, 32, 38, 41

Frequency DistributionsCumulative Distributions

Polygons

Tabulating Numerical Data: Frequency Distributions

Data in ordered array:12, 13, 17, 21, 24, 24, 26, 27, 27, 30, 32, 35, 37, 38, 41, 43, 44, 46, 53, 58

Class Frequency

10 but under 20 3 .15 15

20 but under 30 6 .30 30

30 but under 40 5 .25 25

40 but under 50 4 .20 20

50 but under 60 2 .10 10

Total 20 1 100

RelativeFrequency

Percentage

(continued)

His togram

0

3

6

54

2

00

1

2

3

4

5

67

5 15 25 36 45 55 More

Fre

qu

en

cy

Graphing Numerical Data:The Histogram

Data in ordered array:12, 13, 17, 21, 24, 24, 26, 27, 27, 30, 32, 35, 37, 38, 41, 43, 44, 46, 53, 58

Class Midpoints

No Gaps Between

Bars

Graphing Numerical Data:The Frequency Polygon

Frequency

0

1

2

3

4

5

6

7

5 15 25 36 45 55 More

Data in ordered array:12, 13, 17, 21, 24, 24, 26, 27, 27, 30, 32, 35, 37, 38, 41, 43, 44, 46, 53, 58

Class Midpoints

Cumulative CumulativeClass Frequency % Frequency

10 but under 20 3 15

20 but under 30 9 45

30 but under 40 14 70

40 but under 50 18 90

50 but under 60 20 100

Tabulating Numerical Data:Cumulative Frequency

Data in ordered array:12, 13, 17, 21, 24, 24, 26, 27, 27, 30, 32, 35, 37, 38, 41, 43, 44, 46, 53, 58

Graphing Numerical Data:The Ogive (Cumulative % Polygon)

Data in ordered array:12, 13, 17, 21, 24, 24, 26, 27, 27, 30, 32, 35, 37, 38, 41, 43, 44, 46, 53, 58

O give

0

20

40

60

80

100

120

10 20 30 40 50 60

Tabulating and Graphing Categorical Data: Univariate Data

Categorical Data

Tabulating Data

The Summary Table

Graphing Data

Pie Charts

Pareto DiagramBar Charts

Summary Table(University Revenues)

Revenue Category Amount Percentage (in thousands $)

Patient Services 46.5 42.27

Tuition/fees 32 29.09

Appropriations 15.5 14.09

Grants/Contracts 16 14.55

Total 110 100

Variables are Categorical.

0

5

1 0

1 5

2 0

2 5

3 0

3 5

4 0

4 5

S to c k s B o n d s S a vin g s C D

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

0 1 0 2 0 3 0 4 0 5 0

S to c k s

B o n d s

S a vin g s

C D

Graphing Categorical Data: Univariate Data

Categorical Data

Tabulating Data

The Summary Table

Graphing Data

Pie Charts

Pareto DiagramBar Charts

Bar ChartEnrollment Summary

0 500 1000 1500 2000 2500 3000

Freshmen

Sophomores

Juniors

Seniors

Unclass

Grad

1st Prof

Pie Chart(for a factbook)

Percentages are rounded to the nearest percent.

Students by Classification

Seniors 15%

Sophomores 14%

Juniors

29%

Freshmen42%

Pareto Diagram

Axis for line graph shows

cumulative %

Axis for bar

chart shows % in each

category

0

510

15

2025

30

3540

45

1 2 3 4

0

20

40

60

80

100

120

Tabulating and Graphing Bivariate Categorical Data

•Contingency Tables

•Side by Side Charts

Tabulating Categorical Data: Bivariate Data

Contingency Table: Enrollment by College

Enrollment A&S BUS NRS Total Category

Freshmen 46 55 27 128

Sophomores 32 44 19 95

Juniors 15 20 13 48

Seniors 16 28 7 51

Total 109 147 66 322

Graphing Categorical Data: Bivariate Data

Side by SideChart

0 200 400 600 800 1000 1200

Freshmen

Sophomores

Juniors

Seniors

Unclass

Grad

1st Prof

A&S AH NS

Principles of Graphical Excellence

• Well designed presentation of data that provides:

Substance Statistics Design

• Communicates complex ideas with clarity, precision and efficiency

• Gives the largest number of ideas in the most efficient manner

• Almost always involves several dimensions• Requires telling the truth about the data

Data-Ink Ratio

Data information

Total ink used to print the graphic

“Much of twentieth-century thinking about statistical graphics

has been preoccupied with the question of how some amateurish chart might fool a naive viewer.”

• Using ‘chart junk’

• No relative basis In comparing data Batches

• Compressing the Vertical axis

• No zero point on the

Vertical axis

Errors in Presenting Data

‘Chart Junk’

Good Presentation

1960: $1.00

1970: $1.60

1980: $3.10

1990: $3.80

Minimum Wage Minimum Wage

0

2

4

1960 1970 1980 1990

$

Bad Presentation

Lie Factor

Size of effect shown in graphic

Size of effect in data

No Relative Basis

Good PresentationA’s received by

students.A’s received by

students.

Bad Presentation

0

200

300

FR SO JR SR

Freq.

10%

30%

FR SO JR SR

%

FR = Freshmen, SO = Sophomore, JR = Junior, SR = Senior

Compressing Vertical Axis

Good Presentation

Quarterly Income Quarterly Income

Bad Presentation

0

25

50

Q1 Q2 Q3 Q4

$

0

100

200

Q1 Q2 Q3 Q4

$

No Zero Point on Vertical Axis

Good PresentationMonthly Expenses Monthly Expenses

Bad Presentation

0

39

42

45

J F M A M J

$

36

39

42

45

J F M A M J

$

Graphing the first six months of sales.

36

No Zero Point on Vertical Axis

Good Presentation

Monthly Expenses Monthly Expenses

Bad Presentation

0

20

40

60

J F M A M J

$

36

39

42

45

J F M A M J

$

Graphing the first six months of sales.

Main defense of the lying graphic....

“Well, at least it was approximately correct, we were just trying to show the general

direction of change.”

Presentation Summary

• Organized Numerical Data:• The Ordered Array and Stem-leaf Display

• Tabulated and Graphed Numerical Data• Frequency Distributions: Tables, Histograms, Polygons

• Cumulative Distributions: Tables, the Ogive

Presentation Summary (continued)

• Tabulated and Graphed Univariate Categorical Data: • The Summary Table• Bar and Pie Charts, the Pareto diagram

•Tabulated and Graphed Bivariate Categorical Data: • Contingency Tables• Side by Side charts

• Discussed Graphical Excellence and Common Errors in Presenting Data

“There remain, however, many other consideration in the design

of statistical graphics – not only of efficiency, but also of complexity,

structure, density, and even beauty.”

“Without data, it is anyone’s opinion.”

Author unknown