Data Visualization Models

Coordinating Multiple Visualizations

• Coordinating multiple visualizations of the same architectural model is crucial to ensure consistency with each other and the model.
• Errors in the model are reflected in the visualizations.
• Simplifying assumptions can make coordination easier, such as:
  • Only one visualization operating at a time.
  • Only one tool operating on the model at a time.

Solutions for Coordinating Multiple Visualizations

• Peer-to-Peer Coordination: Each visualization communicates with each other for updates.
  • Has scaling problems.
  • Works best for visualizations known a priori.
• Push-Based Solution: Visualizations actively notify and update themselves whenever the model changes.
  • Best for multiple simultaneous visualizations.
  • Hard to debug, with potential infinite loops and subtle concurrency conditions.

Guidelines for Constructing New Visualizations

• Borrow elements from similar visualizations to leverage existing stakeholder knowledge and improve comprehensibility.
• Be consistent among visualizations, avoiding conflicts with existing visualizations without a good reason.
• Give meaning to each visual aspect of elements, prioritizing parsimony over aesthetics.
• Document the meaning of visualizations.
• Balance traditional and innovative interfaces, using popular notations (e.g., tables).

Models vs. Visualizations

• A model is abstract information, a set of design decisions.
• Visualizations give design decisions form, allowing depiction and interaction with them in different ways.
• Visualizations are often active, being both pictures and tools.

Canonical Visualizations

• Each modeling notation is associated with one or more canonical visualizations.
• This makes it easy to think of a notation and a visualization as the same thing, even though they are not.

Textual Visualizations

• Generally use an ordinary text editor or word processor.
• Some advanced mechanisms available, such as syntax highlighting, static checking, autocomplete, and structural folding.
• Example: XML visualization.

Graphical Visualizations

• Generally use graphical editors with point-and-click interfaces.
• Employ metaphors like scrolling, zooming, and "drill-down".
• Editors have varying levels of awareness for different target notations.
• More exotic editors and interaction mechanisms exist in research, such as 3D editors and "sketching-based" editors.

Hybrid Visualizations

• Many visualizations are text-only.
• Few graphical notations are purely symbolic.
• Text labels and annotations are generally textual as well.
• Some notations incorporate substantial parts that are mostly graphical alongside substantial parts that are mostly or wholly textual.

Evaluating Visualizations

• Scope and Purpose: Consider what the visualization is for.
• Other factors to evaluate: Depiction, Interaction, Dynamism, View Coordination, Aesthetics, and Extensibility.