IEEE VIS 2024 Content: Using Counterfactuals to Improve Causal Inferences From Visualizations

Using Counterfactuals to Improve Causal Inferences From Visualizations

David Borland -

Arran Zeyu Wang -

David Gotz -

Room: Bayshore III

2024-10-17T16:48:00ZGMT-0600Change your timezone on the schedule page
2024-10-17T16:48:00Z
Exemplar figure, described by caption below
A counterfactual subset includes data points from the excluded set that closely resemble those in the included set. Previous research indicates that visualizations comparing the counterfactual subset with the included subset (c) lead to more accurate causal inferences than traditional methods (b). This work will share our vision for how counterfactual concepts developed by the causal inference community can be leveraged to enable the development of more effective visualization technologies.
Fast forward
Keywords

Analytical Models, Correlation, Visual Analytics, Decision Making, Data Visualization, Reliability Theory, Cognition, Inference Algorithms, Causal Inference, Causality, Social Media, Exploratory Analysis, Data Visualization, Visual Representation, Visual Analysis, Visualization Tool, Open Challenges, Interactive Visualization, Assembly Line, Different Subsets Of Data, Visual Analytics Tool, Data Driven Decision Making, Data Quality, Statistical Models, Causal Effect, Visual System, Use Of Social Media, Bar Charts, Causal Model, Causal Graph, Chart Types, Directed Acyclic Graph, Visual Design, Portion Of The Dataset, Causal Structure, Prior Section, Causal Explanations, Line Graph

Abstract

Traditional approaches to data visualization have often focused on comparing different subsets of data, and this is reflected in the many techniques developed and evaluated over the years for visual comparison. Similarly, common workflows for exploratory visualization are built upon the idea of users interactively applying various filter and grouping mechanisms in search of new insights. This paradigm has proven effective at helping users identify correlations between variables that can inform thinking and decision-making. However, recent studies show that consumers of visualizations often draw causal conclusions even when not supported by the data. Motivated by these observations, this article highlights recent advances from a growing community of researchers exploring methods that aim to directly support visual causal inference. However, many of these approaches have their own limitations, which limit their use in many real-world scenarios. This article, therefore, also outlines a set of key open challenges and corresponding priorities for new research to advance the state of the art in visual causal inference.