New Articles Highlight Advances in Counterfactual Research
VACLab researchers published journal articles in February and August highlighting new results from our ongoing work exploring the use of counterfactuals to enhance people’s abilities to draw more meaningful causal inferences from visualizations of data.
Two new papers in the journal Information Visualization describe some of our latest work with counterfactuals to support visual causal inference. Both papers were written by Arran Zeyu Wang, David Borland, and David Gotz.
In the first paper (“An Empirical Study of Counterfactual Visualization to Support Visual Causal Inference”), published in February, we report findings from an empirical study that examined how counterfactuals can improve users’ undersatanding of data in static visualizations. This paper also proposes a conceptual model of causality comprehension that bridges theory from both the causal inference and the visual data communication literature. The article is open access an can be found both on the publisher’s website an in our own library of publications.
The second paper (“A Framework to Improve Causal Inferences from Visualizations Using Counterfactual Operators”), published this month, defines a operator-based computational framework for counterfactuals that formalizes our proposed approach to leveraging counterfactual concepts within iteractive and exploratory visualization systems. This approach is instantiated in our open-source software for counterfactual systems, Co-Op which was announced in this news posting. The article is open access an can be found both on the publisher’s website an in our own library of publications.
Both articles describe work made possible in part by the National Science Foundation under Award #2211845.