Four types of ensemble coding in data visualizations

Abstract

Ensemble coding supports rapid extraction of visual statistics about distributed visual information. Researchers typically study this ability with the goal of drawing conclusions about how such coding extracts information from natural scenes. Here we argue that a second domain can serve as another strong inspiration for understanding ensemble coding: graphs, maps, and other visual presentations of data. Data visualizations allow observers to leverage their ability to perform visual ensemble statistics on distributions of spatial or featural visual information to estimate actual statistics on data. We survey the types of visual statistical tasks that occur within data visualizations across everyday examples, such as scatterplots, and more specialized images, such as weather maps or depictions of patterns in text. We divide these tasks into four categories: identification of sets of values, summarization across those values, segmentation of collections, and estimation of structure. We point to unanswered questions for each category and give examples of such cross-pollination in the current literature. Increased collaboration between the data visualization and perceptual psychology research communities can inspire new solutions to challenges in visualization while simultaneously exposing unsolved problems in perception research.

Assessing Effects of Task and Data Distribution on the Effectiveness of Visual Encodings

“…For summary tasks , dot plots with Q 1 : size outperformed other encodings using position channels , even though size is known to be less effective than position channels for value comparison tasks. This result suggests that size supports effective ensemble coding [SHGF16].…”

“…Much of the prior work focuses on the performance of reading and comparing values encoded by individual visual objects. A more recent trend is the study of tasks involving the perception of distributed (or ensemble) visual information [SHGF16]. Such tasks include perceiving summary values that are not explicitly encoded, such as averages [ACG14, GCNF13] or levels of correlation [HYFC14,KH16].…”

Formalizing Visualization Design Knowledge as Constraints: Actionable and Extensible Models in Draco

“…Casner's BOZ [8] additionally models the low-level perceptual tasks of reading and comparing values. In addition to value tasks, Draco's model includes summary tasks involving aggregate properties of visual ensembles [62]. ShowMe [38] uses heuristic rules to suggest encodings from groups of charts, including trellis plots.…”

“…However, most work on effectiveness focuses on the performance of reading or comparing individual marks in a visualization. Recent work investigates the effects of reading ensembles of visual elements [62]: for example, how users read aggregates [22], distributions, trends, or correlation [24]. Experimental results from Kim et al [30] and Saket et al [51] analyze how effectiveness varies by task.…”

Design Factors for Summary Visualization in Visual Analytics

“…As an example, summary visualizations can use visual encodings that allow analysts to visually estimate features of data distributions from individual datapoints rather than encoding these features directly. Such visual aggregation may enable a more holistic view on data [SHGF16], though we found no summaries explicitly leveraging this strategy.…”