graphical perception of multiple time series
DESCRIPTION
Line graphs have been the visualization of choice for temporal data ever since the days of William Playfair (1759-1823), but realistic temporal analysis tasks often include multiple simultaneous time series. In this work, we explore user performance for comparison, slope, and discrimination tasks for different line graph techniques involving multiple time series. Our results show that techniques that create separate charts for each time series--such as small multiples and horizon graphs--are generally more efficient for comparisons across time series with a large visual span. On the other hand, shared-space techniques--like standard line graphs--are typically more efficient for comparisons over smaller visual spans where the impact of overlap and clutter is reduced.TRANSCRIPT
GraphicalPerception
O F M U LT I P L E T I M E S E R I E S
WaqasJavedBryanMcDonnelNiklasElmqvist
VACCINEVisual Analytics for Command, Control, and Interoperability Environments
A U.S. Department of Homeland Security Center of Excellence1
2
3
Overview• Graphical Perception• Motivation• Contributions• Related Work• Visualization of Multiple Time Series• User Study• Study Result
4
Overview• Graphical Perception• Motivation• Contributions• Related Work• Visualization of Multiple Time Series• User Study• Study Result
5
Graphical Perception
The ability of users to comprehend the visual encoding and thereby decode the information presented in the graph.
The ability of users to comprehend the visual encoding and thereby decode the information presented in the graph, representing multiple time series.
6
Overview• Graphical Perception• Motivation• Contributions• Related Work• Visualization of Multiple Time Series• User Study• Study Result
7
Motivation
• Line graphs: common type of statistical data graphics• Used to visualize temporal data in various domains• Example: finance, politics, science, engineering, and
medicine• Comparison is a common task for time series data• Within same time series, across different time series• Example: Stock analyst, Cardiologist
• Graphical perception of multiple series plays an important role in the success of temporal visualizations
• Effective guidelines are required for designers• Many visualization applications show multiple time series• Find a suitable line graph technique for comparison task
8
Overview• Graphical Perception• Motivation• Contributions• Related Work• Visualization of Multiple Time Series• User Study• Study Result
9
Contributions• We evaluate graphical perception of multiple time series
as a function of different visualization types, under different conditions
• Evaluating the effect of the following conditions
• Visualization type
• Number of series
• Available space
• Task type
10
Overview• Graphical Perception• Motivation• Contributions• Related Work• Visualization of Multiple Time Series• User Study• Study Result
11
Related Work• Graphical perception is not a new research topic • Croxton et al. (1927) compared bar charts with circle
diagrams and pie charts• Cleveland and McGill (1984) formalize the use of graphical
perception for measuring the effectiveness of various graph techniques
• Simkin and Hastie (1987) compared the accuracy of judgment based on comparison and estimation
• Lam et al. (2007) study the differences between low and high-resolution visual representations of line graphs
• Heer et al. (2008) measure the effect of chart size and layering on user performance (horizon graphs)
12
Overview• Graphical Perception• Motivation• Contributions• Related Work• Visualization of Multiple Time Series• User Study• Study Result
13
Line Style Line Color
• We identify five different factors to classify the line graph visualization techniques• Space management• Space per series• Identity • Baseline• Visual clutterShared Space Split Space
Classification Criteria
Available Space = S Available Space = S/N
Common Baseline Individual Baseline
14
Simple Line Graphs (SG)
Space management
Space per series
Identity Baseline Visual Clutter
Shared S Line Common Medium
15
Small Multiples (SM)
Space management
Space per series
Identity Baseline Visual Clutter
Split S/N --- Common Low
16
Stacked Graphs
Space management
Space per series
Identity Baseline Visual Clutter
Shared Proportional Area Previous Medium
17
Horizon Graphs (HG) [Saito 2005]
Image courtesy of [Few 2008]
18
Horizon Graphs (HG)
Space management
Space per series
Identity Baseline Visual Clutter
Split S/N * 2*B --- Common Low
19
Braided Graphs (BG)
20
Braided Graphs (BG)
Space management
Space per series
Identity Baseline Visual Clutter
Shared S Area Common High
21
Overview• Graphical Perception• Motivation• Contributions• Related Work• Visualization of Multiple Time Series• User Study• Study Result
22
Study Hypotheses
H1 Shared-space techniques will perform better for tasks with local visual span
H2 Split-space techniques will perform better for tasks with dispersed visual span
H3 Many concurrent time series will cause decreased performance
H4 Small display space will cause decreased performance
23
Tasks
• Maximum: local comparison across all time series
• Discrimination: dispersed comparison of time series
• Slope: dispersed rate estimation across all time series
24
Study Design• Visualization type (V)• SG, BG, SM, HG
• Tasks (T)• Maximum, Slope, Discrimination
• Number of time series (N)• 2, 4, 8
• Total Chart Size (S)• 48 px (small), 96 px (medium), 192 px (large)
25
Overview• Graphical Perception• Motivation• Contributions• Related Work• Visualization of Multiple Time Series• User Study• Study Result
26
Correctness vs. Visualization type
27
Completion time vs. Number of time series
28
Completion time vs. Chart size
29
Completion time vs. Visualization type
Discrimination task Maximum task Slope task
30
Summary of Findings• Shared-space techniques (SG and BG) were faster than
splits-space techniques for Maximum (H1 )• Split-space techniques (SM and HG) were faster than
shared-space techniques for Discrimination (H2 )• The Slope task, with dispersed visual span, was special
—SM and SG were fastest here• Higher numbers of concurrent time series caused
decreased correctness and increased completion time (H3 )
• Decreased display space allocation had a negative impact on correctness, but had little effect on time (partially confirming H4)
31
Conclusion
• I have presented results from a user study on the graphical perception of multiple simultaneous time series
• Results from our experiment indicate that • Superimposed/shared space line graph techniques
work best for local tasks• Juxtaposed/split space techniques work best for
dispersed ones
GraphicalPerception
O F M U LT I P L E T I M E S E R I E S
Waqas JavedE-mail: [email protected]
Website: http://web.ics.purdue.edu/~wjaved/
VACCINEVisual Analytics for Command, Control, and Interoperability Environments
A U.S. Department of Homeland Security Center of Excellence
Thanks!Acknowledgments:
This research was partially funded by Google, Inc., under the project “Multi-Focus Interaction for Time-Series Visualization”.
33
More Information
• Online versionhttps://engineering.purdue.edu/~elm/projects/gvis/
• Pivot website:https://engineering.purdue.edu/pivot/
• Pivot on Facebookhttp://www.facebook.com/#!/
pages/PivotLab/131505430222567