Teru Teru Bōzu: Defensive Raincloud Plots

Abstract

Univariate visualizations like histograms, rug plots, or box plots provide concise visual summaries of distributions. However, each individual visualization may fail to robustly distinguish important features of a distribution, or provide sufficient information for all of the relevant tasks involved in summarizing univariate data. One solution is to juxtapose or superimpose multiple univariate visualizations in the same chart, as in Allen et al.'s [APW*19] “raincloud plots.” In this paper I examine the design space of raincloud plots, and, through a series of simulation studies, explore designs where the component visualizations mutually “defend” against situations where important distribution features are missed or trivial features are given undue prominence. I suggest a class of “defensive” raincloud plot designs that provide good mutual coverage for surfacing distributional features of interest.

On The Reasonable Effectiveness of Relational Diagrams: Explaining Relational Query Patterns and the Pattern Expressiveness of Relational Languages

“…Our choice of visualization is a variant of Raincloud plots [5] and is inspired from recent work in the visualization literature [24] discussing various ways to juxtapose multiple visualizations ("clouds + rain + lightning") in the same chart for increasing information content. In that framework, each of our charts consists of (𝑖) density plots that show an overview of the shape of the distribution (the "cloud"), (𝑖𝑖) unjittered dot plots that show the raw data (the "rain": here we deviate from [24] in using triangles instead of circles which, in our opinion, are more easily countable due to their visible vertices), and (𝑖𝑖𝑖) 95% confidence intervals that provide summary statistics (the "lightning"). Furthermore, whenever we compare alternative modalities ("repeated measures"), we also use ( 4) paired plots with lines connecting summary statistics and/or raw data.…”

“…relational division and is shown as Relational Diagrams, SQL * , RA * , and Datalog * in Figs 24. and 25, and will also be re-used in Example 18.…”

Visualization Guardrails: Designing Interventions Against Cherry-Picking in Interactive Data Explorers

“…The space of ways to show the distribution of contextual data is large and the choice of a specific visual encoding should depend on the features of the underlying distribution, as well as the expected audience's visual literacy, as these types of charts are typically less common in visualizations for general public. Correll provides a helpful analysis of advantages and disadvantages of distribution visualizations as well as their combinations as raincloud plots [11]. The example in Figure 1 as well as our prototype employ a vertical strip plot, but other designs we have considered include a box plot, a violin chart, or an inverted histogram (shown in the example on right in Figure 4).…”