Text in Visualization: Extending the Visualization Design Space

R Brath
2018
This thesis is a systematic exploration and expansion of the design space of data visualization specifically with regards to text. A critical analysis of text in data visualizations reveals gaps in existing frameworks and the use of text in practice. A cross-disciplinary review across fields such as typography, cartography and technical applications yields typographic techniques to encode data into text and provides the scope for the expanded design space. Mapping new attributes, techniques and
more » ... considerations back to well understood visualization principles organizes the design space of text in visualization. This design space includes: 1) text as a primary data type literally encoded into alphanumeric glyphs, 2) typographic attributes, such as bold and italic, capable of encoding additional data onto literal text, 3) scope of mark, ranging from individual glyphs, syllables and words; to sentences, paragraphs and documents, and 4) layout of these text elements applicable most known visualization techniques and text specific techniques such as tables. This is the primary contribution of this thesis (Part A and B). Then, this design space is used to facilitate the design, implementation and evaluation of new types of visualization techniques, ranging from enhancements of existing techniques, such as, extending scatterplots and graphs with literal marks, stem & leaf plots with multivariate glyphs and broader scope, and microtext line charts; to new visualization techniques, such as, multivariate typographic thematic maps; text formatted to facilitate skimming; and proportionally encoding quantitative values in running text – all of which are new contributions to the field (Part C). Finally, a broad evaluation across the framework and the sample visualizations with cross-discipline expert critiques and a metrics based approach reveals some concerns and many opportunities pointing towards a breadth of future research work now possible with this new framework. (Part D and E).
doi:10.18744/pub.002743 fatcat:3fidoayq45d2dpfb6jtl2ftcoi