VisDock: A Toolkit for Cross-cutting Interactions in Visualization
Jungu "Joe" Choi, Yuet Ling Wong, Eli Fisher
University of Maryland, College Park, MD
Deok Gun Park, Niklas Elmqvist
Purdue University, W. Lafayette, IN
Motivation
Cross-cutting interactions for visualization
interaction techniques that are common to a range of visual representations, data and tasks
For Developers
- Provides a guideline
For end-users
- Reduces learning effort
How?
- Identify common interactions
- Develop a toolkit
- Evaluate
1. Identifying common interactions
- Any dataset
- Any visualization
- Concrete
- Direct manipulation
Design Criteria
Graphics Software
Interactive Visualizations
Cross-cutting Interaction for
Visualization
Cross-cutting Interactions for Visualization
- Selection
Cross-cutting Interactions for Visualization
- Selection
2. Query Management
Cross-cutting Interactions for Visualization
1. Selection
2. Query Management
3. Navigation
Cross-cutting Interactions for Visualization
1. Selection
2. Query Management
3. Navigation
4. Annotation
Cross-cutting Interactions for Visualization
- Selection
- Query Management
- Navigation
- Annotation
3. VisDock
Visualization built by Derek Watkins
http://bl.ock.org/dwtkns
VisDock
a mix-in JavaScript library
- to import cross-cutting interactions in any SVG-oriented visualizations
- to connect the interactive events to user-defined functions
Minimal coding for Developers
Still you need to write event handler.
Familiar UI for Users
4. Evaluation
VisDock-enabled
Visualization
Developer
End-User
VisDock Evaluation I
- Informal evaluation study with dozen programmers
Evaluation I Results
- VisDock successfully imported
- VisDock Event Handler partial implementation
VisDock Evaluation II
- 11 paid volunteers to use VisDock-integrated MovieViz visualization.
VisDock Evaluation II
- What is MovieViz?
- Complex visualization with ~ 1000 multi-variate movie and movie-goers' data
Evaluation II Results
- Most participants (9 out of 10) successfully accomplished all objectives.
- All participants queried data using selection tools
- About 6.5 insightful comments per participant
Conclusion
Contributions
- Concept of cross-cutting interactions
- Design of Cross-cutting interactions for Visualization
- Visdock as an open-source toolkit
Check out our Git Repository
http://visdockhub.github.io/NewVisDock
Team members
Jungu "Joe" Choi
choi88@purdue.edu
Deok Gun Park
intuinno@umd.edu
Niklas Elmqvist
elm@umd.edu
