ABSTRACT
Preattentive patterns in text can be created using visual and typographic attributes. Further, visualizations that organize inventories of data or communicate explanations may require detailed textual data directly in the visualization. However, the current conceptualization of the visualization encoding pipeline maps structured data to visual attributes drawn as marks and plotted on a layout, without regards for text. This pipeline can be naively extended by preprocessing raw text into structured data. Instead, as informed by historic examples, each stage in the visualization pipeline is extended: (1) structured data add a literal data type; (2) visual attributes are extended to typographic attributes such as alphanumerics, weight, case, and typeface; (3) marks are adapted such that glyphs and words are point marks, phrases and sentences are line marks, and longer text are area marks; and (4) layouts are extended to text-specific layouts such as prose and tables.
