Vega-Lite: A Grammar of Interactive Graphics

Area of Science:

• Computer Science
• Data Visualization
• Human-Computer Interaction

Background:

• Traditional grammars of graphics facilitate data visualization but often lack robust interaction design capabilities.
• Specifying interactive semantics in data visualization can be complex and time-consuming.

Purpose of the Study:

• To introduce Vega-Lite, a high-level grammar for the rapid specification of interactive data visualizations.
• To present a novel grammar of interaction integrated with a grammar of graphics.
• To enable users to define interactive semantics through composable selections.

Main Methods:

• Vega-Lite combines a grammar of graphics with a novel grammar of interaction.
• Interactive semantics are specified by composing "selections," which define event processing, points of interest, and predicate functions.
• Selections parameterize visual encodings and drive conditional logic, with an automated compiler synthesizing data flow and event handling logic.

Main Results:

• Vega-Lite allows for the succinct specification of customized interaction methods.
• Common interactive techniques such as panning, zooming, and linked selection are demonstrated.
• The system automatically generates necessary data flow and event handling logic, which can be customized.

Conclusions:

• Vega-Lite offers a powerful and efficient approach to designing interactive data visualizations.
• The selection-based interaction model simplifies the creation of complex interactive behaviors.
• Vega-Lite enhances the expressiveness and usability of data visualization tools.