The document describes an exploratory study on data sketching for visual representation. Researchers conducted sessions where participants were asked to sketch representations of behavior-situation data on paper in any way they wished. Participants created a variety of representations ranging from numeric plots and matrices to more abstract pictorial designs. Analysis of the 35 sketches created a continuum from more numerically-focused and countable designs to more abstract visualizations. Common representation types included dot plots, matrices, bar charts, line graphs and parallel coordinates for more data-driven sketches and pictorial, graph-like and ranked list designs for more abstract sketches.

1. AN EXPLORATORY STUDY
OF DATA SKETCHING
FOR VISUAL REPRESENTATION
Jagoda Walny, Samuel Huron, and Sheelagh Carpendale
InnoVis & Interactions Lab, University of Calgary
EuroVis 2015, Cagliari, Italy

2. Why data sketching?

3. ➡Low barrier to entry
3

4. ➡Low barrier to entry
➡Rapid
3

5. ➡Low barrier to entry
➡Rapid
➡Freeform
3

6. Data Sketching Study

7. Data Sketching Study
“Represent the data on the blank sheets of paper
in any way you wish”

8. Data Sketching Study
9F
13M
3 sessions: 7 / 8 / 7
5

9. Procedure
➡Scripted
➡Planned answers
➡Questionnaire
6

10. “Please describe what you learned or found
interesting about this data during the session.
(there are no wrong answers)”
7

11. Dataset
Run Talk Kiss Write Eat Sleep Mumble Read Fight Belch Argue Jump Cry Laugh Shout
Class 2.52 6.21 2.10 8.17 4.23 3.60 3.62 7.27 1.21 1.77 5.33 1.79 2.21 6.23 1.94
Date 5.00 8.56 8.73 3.62 7.79 3.77 3.12 2.88 3.58 2.23 4.50 4.42 3.04 8.00 3.79
Bus 1.44 8.08 4.27 4.87 5.48 7.04 5.17 7.17 1.52 2.15 4.17 3.12 3.08 7.10 3.00
Family dinner 2.56 8.52 4.92 2.58 8.44 2.29 2.54 3.96 1.67 2.50 3.25 2.29 3.21 7.13 1.96
Park 7.94 8.42 7.71 7.00 8.13 5.63 5.40 7.77 3.06 5.00 5.06 7.42 5.21 8.10 6.92
Church 1.38 3.29 2.38 2.85 1.38 1.77 3.52 3.58 0.62 1.42 1.92 1.71 3.13 2.60 1.33
Job interview 1.94 8.46 1.08 4.85 1.73 0.75 1.31 2.48 1.04 1.21 1.83 1.48 1.37 5.88 1.65
Sidewalk 5.58 8.19 4.75 3.38 4.83 1.46 4.96 4.81 1.46 2.81 4.08 3.54 3.71 7.40 4.88
Movies 2.46 4.98 6.21 2.73 7.48 4.08 4.13 1.73 1.37 2.58 1.71 2.31 7.15 7.94 2.42
Bar 1.96 8.25 5.17 5.38 7.67 2.90 6.21 4.71 1.90 5.04 4.31 3.75 3.44 8.23 4.13
Elevator 1.63 7.40 4.79 3.04 5.10 1.31 5.12 4.48 1.58 2.54 2.58 2.12 3.48 6.77 1.73
Restroom 2.83 7.25 2.81 3.46 2.35 2.83 5.04 4.75 1.77 5.12 3.48 3.65 4.79 5.90 3.52
Own room 6.15 8.58 8.52 8.29 7.94 8.85 7.67 8.58 4.25 6.81 7.52 6.73 8.00 8.17 6.44
Dorm lounge 4.40 7.88 6.54 7.73 7.19 6.08 5.50 8.56 2.40 4.00 4.88 4.58 3.88 7.75 3.60
Football game 4.12 8.08 5.08 4.56 8.04 2.98 5.23 3.69 2.04 3.85 4.98 7.12 4.31 7.90 7.94
9 = "The behavior is extremely appropriate in this situation."
Behavior
Situation
Mean Appropriateness Ratings for 225 Behavior-Situation Combinations
Note:
0 = "The behavior is extremely inappropriate in this situation." Price R. H., Bouffard D. L.: Behavioral appropriateness and situational constraint as
dimensions of social behavior. Journal of Personality and Social Psychology (1974).
8

12. Dataset
Run Talk Kiss Write Eat Sleep Mumble Read Fight Belch Argue Jump Cry Laugh Shout
Class 2.52 6.21 2.10 8.17 4.23 3.60 3.62 7.27 1.21 1.77 5.33 1.79 2.21 6.23 1.94
Date 5.00 8.56 8.73 3.62 7.79 3.77 3.12 2.88 3.58 2.23 4.50 4.42 3.04 8.00 3.79
Bus 1.44 8.08 4.27 4.87 5.48 7.04 5.17 7.17 1.52 2.15 4.17 3.12 3.08 7.10 3.00
Family dinner 2.56 8.52 4.92 2.58 8.44 2.29 2.54 3.96 1.67 2.50 3.25 2.29 3.21 7.13 1.96
Park 7.94 8.42 7.71 7.00 8.13 5.63 5.40 7.77 3.06 5.00 5.06 7.42 5.21 8.10 6.92
Church 1.38 3.29 2.38 2.85 1.38 1.77 3.52 3.58 0.62 1.42 1.92 1.71 3.13 2.60 1.33
Job interview 1.94 8.46 1.08 4.85 1.73 0.75 1.31 2.48 1.04 1.21 1.83 1.48 1.37 5.88 1.65
Sidewalk 5.58 8.19 4.75 3.38 4.83 1.46 4.96 4.81 1.46 2.81 4.08 3.54 3.71 7.40 4.88
Movies 2.46 4.98 6.21 2.73 7.48 4.08 4.13 1.73 1.37 2.58 1.71 2.31 7.15 7.94 2.42
Bar 1.96 8.25 5.17 5.38 7.67 2.90 6.21 4.71 1.90 5.04 4.31 3.75 3.44 8.23 4.13
Elevator 1.63 7.40 4.79 3.04 5.10 1.31 5.12 4.48 1.58 2.54 2.58 2.12 3.48 6.77 1.73
Restroom 2.83 7.25 2.81 3.46 2.35 2.83 5.04 4.75 1.77 5.12 3.48 3.65 4.79 5.90 3.52
Own room 6.15 8.58 8.52 8.29 7.94 8.85 7.67 8.58 4.25 6.81 7.52 6.73 8.00 8.17 6.44
Dorm lounge 4.40 7.88 6.54 7.73 7.19 6.08 5.50 8.56 2.40 4.00 4.88 4.58 3.88 7.75 3.60
Football game 4.12 8.08 5.08 4.56 8.04 2.98 5.23 3.69 2.04 3.85 4.98 7.12 4.31 7.90 7.94
9 = "The behavior is extremely appropriate in this situation."
Behavior
Situation
Mean Appropriateness Ratings for 225 Behavior-Situation Combinations
Note:
0 = "The behavior is extremely inappropriate in this situation." Price R. H., Bouffard D. L.: Behavioral appropriateness and situational constraint as
dimensions of social behavior. Journal of Personality and Social Psychology (1974).
8

13. Dataset
Run Talk Kiss Write Eat Sleep Mumble Read Fight Belch Argue Jump Cry Laugh Shout
Class 2.52 6.21 2.10 8.17 4.23 3.60 3.62 7.27 1.21 1.77 5.33 1.79 2.21 6.23 1.94
Date 5.00 8.56 8.73 3.62 7.79 3.77 3.12 2.88 3.58 2.23 4.50 4.42 3.04 8.00 3.79
Bus 1.44 8.08 4.27 4.87 5.48 7.04 5.17 7.17 1.52 2.15 4.17 3.12 3.08 7.10 3.00
Family dinner 2.56 8.52 4.92 2.58 8.44 2.29 2.54 3.96 1.67 2.50 3.25 2.29 3.21 7.13 1.96
Park 7.94 8.42 7.71 7.00 8.13 5.63 5.40 7.77 3.06 5.00 5.06 7.42 5.21 8.10 6.92
Church 1.38 3.29 2.38 2.85 1.38 1.77 3.52 3.58 0.62 1.42 1.92 1.71 3.13 2.60 1.33
Job interview 1.94 8.46 1.08 4.85 1.73 0.75 1.31 2.48 1.04 1.21 1.83 1.48 1.37 5.88 1.65
Sidewalk 5.58 8.19 4.75 3.38 4.83 1.46 4.96 4.81 1.46 2.81 4.08 3.54 3.71 7.40 4.88
Movies 2.46 4.98 6.21 2.73 7.48 4.08 4.13 1.73 1.37 2.58 1.71 2.31 7.15 7.94 2.42
Bar 1.96 8.25 5.17 5.38 7.67 2.90 6.21 4.71 1.90 5.04 4.31 3.75 3.44 8.23 4.13
Elevator 1.63 7.40 4.79 3.04 5.10 1.31 5.12 4.48 1.58 2.54 2.58 2.12 3.48 6.77 1.73
Restroom 2.83 7.25 2.81 3.46 2.35 2.83 5.04 4.75 1.77 5.12 3.48 3.65 4.79 5.90 3.52
Own room 6.15 8.58 8.52 8.29 7.94 8.85 7.67 8.58 4.25 6.81 7.52 6.73 8.00 8.17 6.44
Dorm lounge 4.40 7.88 6.54 7.73 7.19 6.08 5.50 8.56 2.40 4.00 4.88 4.58 3.88 7.75 3.60
Football game 4.12 8.08 5.08 4.56 8.04 2.98 5.23 3.69 2.04 3.85 4.98 7.12 4.31 7.90 7.94
9 = "The behavior is extremely appropriate in this situation."
Behavior
Situation
Mean Appropriateness Ratings for 225 Behavior-Situation Combinations
Note:
0 = "The behavior is extremely inappropriate in this situation." Price R. H., Bouffard D. L.: Behavioral appropriateness and situational constraint as
dimensions of social behavior. Journal of Personality and Social Psychology (1974).
8

14. AUDIENCE PARTICIPATION!

15. j.mp/datasketching
AUDIENCE PARTICIPATION!

16. Analysis
35 Representations + Dataset sheets + Questionnaires
➡ Qualitative coding
➡ Collaborative
10

17. Continuum:
Numeracy to Abstractness

18. COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
12

19. COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
12

20. COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
12

21. COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
13
Countable, Dot Plots & Matrices

22. COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
14
Countable, Dot Plots & Matrices

23. COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
14
Countable, Dot Plots & Matrices

24. COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
15
Bar Charts, Line Graphs, Parallel Coordinates

25. 16
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Bar Charts, Line Graphs, Parallel Coordinates

26. 17
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Ranked Lists

27. 18
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Graph-like Representations & Venn Diagrams

28. 19
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Graph-like Representations & Venn Diagrams

29. 19
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Graph-like Representations & Venn Diagrams

30. 20
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Pictorial Representations

31. 21

32. 22
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Pictorial Representations

33. 23
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Pictorial Representations

34. 23
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
Pictorial Representations

35. Spectrum of Data
Reports
“Please describe what you learned or
found interesting about this data during
the session.
(there are no wrong answers)”

36. A B C D E F
Information Intrinsic to Dataset
25

37. A B C D E F
Information Intrinsic to Dataset
• A. Individual values
• e.g. “fighting in church is inappropriate”
25

38. A B C D E F
Information Intrinsic to Dataset
• A. Individual values
• e.g. “fighting in church is inappropriate”
• B. Summarized rows or columns
• e.g. “there aren’t many behaviours appropriate in church”
25

39. A B C D E F
Information Intrinsic to Dataset
• A. Individual values
• e.g. “fighting in church is inappropriate”
• B. Summarized rows or columns
• e.g. “there aren’t many behaviours appropriate in church”
• C. Compared two rows or columns
• “Date and own room have the similar rating for ‘kiss’. Other ratings in
these two situations are close to each other.”
25

40. A B C D E F
Dataset-level Trends and Comparisons
26

41. A B C D E F
Dataset-level Trends and Comparisons
• Compare three or more rows/columns
26

42. A B C D E F
Dataset-level Trends and Comparisons
• Compare three or more rows/columns
• Group items by value
26

43. A B C D E F
Dataset-level Trends and Comparisons
• Compare three or more rows/columns
• Group items by value
• Make global comparisons
26

44. A B C D E F
Dataset-level Trends and Comparisons
• Compare three or more rows/columns
• Group items by value
• Make global comparisons
• E.g. “Several situations which have a similar rate for one specific behavior
tend to be similar for other behaviors.”
26

45. A B C D E F
Dataset-level Trends and Comparisons
• Compare three or more rows/columns
• Group items by value
• Make global comparisons
• E.g. “Several situations which have a similar rate for one specific behavior
tend to be similar for other behaviors.”
• E.g. Binning: “completely appropriate”, “somewhat appropriate”,
“highly inappropriate”
26

46. A B C D E F
Including Extrinsic Information
27

47. A B C D E F
Including Extrinsic Information
• Classify
• e.g. “comfortable”, “safe”, “aggressive”
27

48. A B C D E F
Including Extrinsic Information
• Classify
• e.g. “comfortable”, “safe”, “aggressive”
• Compare to expectations
• “mumbling + talking diverged more than expected.”
27

49. A B C D E F
Including Extrinsic Information
• Classify
• e.g. “comfortable”, “safe”, “aggressive”
• Compare to expectations
• “mumbling + talking diverged more than expected.”
• Explain in domain context
• “people care a lot in job interviews”
27

50. A B C D E F
Analytic Potential
28

51. A B C D E F
Analytic Potential
• Hypotheses or conjectures about the reasons behind the
values in the dataset
28

52. A B C D E F
Analytic Potential
• Hypotheses or conjectures about the reasons behind the
values in the dataset
• E.g. “it appears the park might be the same as one's own
room... anonymity? “
28

53. A B C D E F
Analytic Potential
• Hypotheses or conjectures about the reasons behind the
values in the dataset
• E.g. “it appears the park might be the same as one's own
room... anonymity? “
• E.g. “I found out that there seem to be more women in
the dataset than men because most inappropriate
behaviours to men (i.e. Talking in the restroom) is still
above 5. ”
28

54. Integration of
Results

55. A B C D FE
EXP

57. EXP

58. A B C D FE
EXP

59. A B C D FE
EXP
1 2 3 4

60. EXP

61. A B C D FE
EXP

62. A B C D FE
EXP
1 2 3 4

63. A B C D FE 1 2 3 4
EXP

64. A B C D FE 1 2 3 4
EXP

65. A B C D FE 1 2 3 4
EXP
COUNTABLEDOTPLOTS&
MATRICES
BARCHARTSLINEGRAPH&
PARALLELCOORDS
GRAPHLIKEVENNPICTORIALRANKEDLIST
NUMERICABSTRACT

66. A B C D FE 1 2 3 4
EXP

67. A B C D FE 1 2 3 4
EXP

68. Levels of
Data Description
36

69. Levels of Data Description

70. Levels of Data Description

71. A B C D E F
Levels of Data Description

72. A B C D E F
1
1. Value-level of Data Description
38

74. A B C D E F
1
2. Dimension-level of Data Description
40

75. A B C D E F
1
2. Dimension-level of Data Description
2
40

76. 41

77. A B C D E F
1
2
3. Global level of Data Description
42

78. 3
A B C D E F
1
2
3. Global level of Data Description
42

79. 43

80. A B C D E F
1
2
3
4. External level of Data Description
44

81. 4
A B C D E F
1
2
3
4. External level of Data Description
44

82. 45

83. 46

84. A B C D E F
1
2
3
4
Levels of Data Description
47

85. Going forward…

86. ➡ Viewing representations in terms of levels
of data description
49

87. ➡ Viewing representations in terms of levels
of data description
➡ Understanding advantages of sketching
for representation and data understanding
49

88. COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
50

89. A B C D E F
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
50

90. 1
2
3
4
A B C D E F
COUNTABLE DOT PLOTS &
MATRICES
BARCHARTS LINE GRAPH &
PARALLEL COORDS
GRAPH LIKE VENN PICTORIALRANKED LIST
NUMERIC ABSTRACT
50

92. Thank you:
An Exploratory Study of Data Sketching for
Visual Representation
Jagoda Walny - jkwalny@ucalgary.ca
Samuel Huron - samuel.huron@cybunk.com
Sheelagh Carpendale - sheelagh@ucalgary.ca
PROJECT WEBSITE: j.mp/datasketching

93. PROJECT WEBSITE: j.mp/datasketching
Send us your sketches! @jawalnut • jkwalny@ucalgary.ca