Lowering the Threshold of Visualization Design

Consuming data visualizations has become mainstream, with people and organizations embracing visualizations to record, analyze, and communicate data. However, designing effective visualizations remains difficult, as it requires a cross-cutting set of expertise. For example, designers need storytelling expertise to select visual forms that convey both the semantics and connotations of the data, design expertise to ensure visual and interactive elements are perceptually sound, and technical expertise to implement and publish the resultant visualization. In this talk, Arvind Satyanarayan presents two projects that begin to lower the threshold of custom visualization design by reducing necessary technical expertise. The first project, Lyra, is a new visualization design environment (VDE) that enables direct-manipulation authoring of visualizations. Data is imported and transformed visually, and drag-and-drop operations bind data values to the properties of graphical primitives. As a result, designers can create highly customized visualizations without any programming. Since its alpha release in March 2014, approximately 2,000 users have used Lyra each month and reported it as being an effective prototyping and teaching tool. The second project, Reactive Vega, formulates a grammar of interaction design. Rather than constructing imperative event-handling callbacks, Reactive Vega introduces a set of primitives that can be composed declaratively in order to have input events trigger visual changes. As a result designers need only focus on specifying their interaction technique and leave the library to manage the complexity of propagating events and changes.