Interaction Techniques for Visual Exploration Using Embedded Word-Scale Visualizations

Abstract

We describe a design space of view manipulation interactions for small data-driven contextual visualizations (word-scale visualizations). These interaction techniques support an active reading experience and engage readers through exploration of embedded visualizations whose placement and content connect them to specific terms in a document. A reader could, for example, use our proposed interaction techniques to explore word-scale visualizations of stock market trends for companies listed in a market overview article. When readers wish to engage more deeply with the data, they can collect, arrange, compare, and navigate the document using the embedded word-scale visualizations, permitting more visualizationcentric analyses. We support our design space with a concrete implementation, illustrate it with examples from three application domains, and report results from two experiments. The experiments show how view manipulation interactions helped readers examine embedded visualizations more quickly and with less scrolling and yielded qualitative feedback on usability and future opportunities.

Charagraph: Interactive Generation of Charts for Realtime Annotation of Data-Rich Paragraphs

“…Victor's definition of active reading focuses on interactive documents to explore, build intuition, make discoveries, and reinforce or challenge knowledge. This idea has been applied to different contexts, such as transparent statistical reports [32] and visual exploration [46] (see Hohman et al [59] for a review).…”

“…Otherwise, it has two interaction contexts based on the position of the pointer: in-text and in-chart. Previous work on reading behaviours with interactive documents [25,46,160] suggests that readers use in-text interactions to follow the document narrative and in-chart interactions to explore and engage with the data.…”

Comparative Layouts Revisited: Design Space, Guidelines, and Future Directions

“…5F-G): Although superposed line charts overall showed superior performance than chart-wise versions, comparing values between different time points for different trends made the superposed line charts significantly ineffective compared to the chart-wise one (see the "Discrimination" task). If the comparison targets can be predefined, they can be aligned in the visualization by default, or interactive designs that allow people to use various alignments on demand can be adopted [20].…”

Does Interaction Improve Bayesian Reasoning with Visualization?

“…Others have endeavored to better explain why interaction is useful to visualization from a cognitive processing standpoint [27,34]. In addition to theorizing and categorizing interaction, work has been done designing novel interaction techniques, for example [6,18,26,44], and identifying how visualization designers can leverage interaction to learn about users and create customized visualizations [5,13].…”