The Garden of Forking Paths in Visualization: A Design Space for Reliable Exploratory Visual Analytics : Position Paper

Pu, Xiaoying; Kay, Matthew

doi:10.1109/beliv.2018.8634103

Cited by 25 publications

(18 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An important question for system designers is how to help users conduct safe visual analysis [101]. Analysts can deceive themselves with statistical traps [65], visualization hallucinations [42], or false graphical inferences [93].…”

Section: Validation In Visualmentioning

confidence: 99%

Integrated Visualization Editing via Parameterized Declarative Templates

McNutt

Chugh

2021

Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

Interfaces for creating visualizations typically embrace one of several common forms. Textual specifcation enables fne-grained control, shelf building facilitates rapid exploration, while chart choosing promotes immediacy and simplicity. Ideally these approaches could be unifed to integrate the user-and usage-dependent benefts found in each modality, yet these forms remain distinct.We propose parameterized declarative templates, a simple abstraction mechanism over JSON-based visualization grammars, as a foundation for multimodal visualization editors. We demonstrate how templates can facilitate organization and reuse by factoring the more than 160 charts that constitute Vega-Lite's example gallery into approximately 40 templates. We exemplify the pliability of abstracting over charting grammars by implementing-as a template-the functionality of the shelf builder Polestar (a simulacra of Tableau) and a set of templates that emulate the Google Sheets chart chooser. We show how templates support multimodal visualization editing by implementing a prototype and evaluating it through an approachability study. CCS CONCEPTS• Human-centered computing → Visualization systems and tools; Graphical user interfaces.

show abstract

Section: Validation In Visualmentioning

confidence: 99%

Integrated Visualization Editing via Parameterized Declarative Templates

McNutt

Chugh

2021

Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

show abstract

“…We followed a process of model expansion with regularization [53,65]. It allowed us to understand how each predictor affects the model, to capture more variance in the data while reducing overfitting, and to explore the effects of secondary variables.…”

Section: Bayesian Multilevel (Hierarchical) Modelingmentioning

confidence: 99%

Rethinking the ranks of visual channels

McColeman¹,

Yang²,

Brady³

et al. 2021

Preprint

View full text Add to dashboard Cite

Data can be visually represented using visual channels like position, length or luminance. An existing ranking of these visual channels is based on how accurately participants could report the ratio between two depicted values. There is an assumption that this ranking should hold for different tasks and for different numbers of marks. However, there is surprisingly little existing work that tests this assumption, especially given that visually computing ratios is relatively unimportant in real-world visualizations, compared to seeing, remembering, and comparing trends and motifs, across displays that almost universally depict more than two values.To simulate the information extracted from a glance at a visualization, we instead asked participants to immediately reproduce a set of values from memory after they were shown the visualization. These values could be shown in a bar graph (position (bar)), line graph (position (line)), heat map (luminance), bubble chart (area), misaligned bar graph (length), or `wind map' (angle). With a Bayesian multilevel modeling approach, we show how the rank positions of visual channels shift across different numbers of marks (2, 4 or 8) and for bias, precision, and error measures. The ranking did not hold, even for reproductions of only 2 marks, and the new probabilistic ranking was highly inconsistent for reproductions of different numbers of marks. Other factors besides channel choice had an order of magnitude more influence on performance, such as the number of values in the series (e.g., more marks led to larger errors), or the value of each mark (e.g., small values were systematically overestimated). Every visual channel was worse for displays with 8 marks than 4, consistent with established limits on visual memory. These results point to the need for a body of empirical studies that move beyond two-value ratio judgments as a baseline for reliably ranking the quality of a visual channel, including testing new tasks (detection of trends or motifs), timescales (immediate computation, or later comparison), and the number of values (from a handful, to thousands)

show abstract

“…Each of these decisions could result in a dramatically different final analysis or visualization. This "garden of forking paths" problem [54] for visual analytics is often portrayed as an issue of reliability or replicability of findings. An emerging challenge in visual analytics is therefore how to capture and visualize the data flow that led to a particular experimental outcome [29], or the visualize the robustness of a conclusion across a "multiverse" of different analytical paths [24,37,47].…”

Section: Sensitivity Analyses Of Visual Analyticsmentioning

confidence: 99%

“…El-Assady et al [26] is an example of what this sort of guidance might look like in a text analytics context: the system proposes optimizations based on speculative execution of particular branches of the parameter space. The direct exposure of the "forking paths" [54] or "multiverse" [47] of analyses could allow users to take ownership of the model while still being cognizant of changes to model structure or performance. A human-in-the-loop system need not simulate the entire complex parameter space, but, as in Lee et al 's [42] "cruise control" metaphor, be guided by the user to particular areas, and then perform local exploration of parameter space to find areas with the best outcomes.…”

Section: Baseline Runmentioning

confidence: 99%