Beyond Correlation: Incorporating Counterfactual Guidance to Better Support Exploratory Visual Analysis

TL;DR

This work integrated counterfactual guidance into an exploratory visual analytics system, and using a synthetically generated ground-truth causal dataset, conducted a comparative user study and evaluated to what extent counterfactual guidance can help lead users to more precise visual causal inferences.

Abstract

Providing effective guidance for users has long been an important and challenging task for efficient exploratory visual analytics, especially when selecting variables for visualization in high-dimensional datasets. Correlation is the most widely applied metric for guidance in statistical and analytical tools, however a reliance on correlation may lead users towards false positives when interpreting causal relations in the data. In this work, inspired by prior insights on the benefits of counterfactual visualization in supporting visual causal inference, we propose a novel, simple, and efficient counterfactual guidance method to enhance causal inference performance in guided exploratory analytics based on insights and concerns gathered from expert interviews. Our technique aims to capitalize on the benefits of counterfactual approaches while reducing their complexity for users. We integrated counterfactual guidance into an exploratory visual analytics system, and using a synthetically generated ground-truth causal dataset, conducted a comparative user study and evaluated to what extent counterfactual guidance can help lead users to more precise visual causal inferences. The results suggest that counterfactual guidance improved visual causal inference performance, and also led to different exploratory behaviors compared to correlation-based guidance. Based on these findings, we offer future directions and challenges for incorporating counterfactual guidance to better support exploratory visual analytics.

Figures (4)

• Figure 1: A comparison between CoFactkaul2021improving (a) and our interface (b). The red indicates the visualized counterfactual information shown to guide user analysis in each interface, demonstrating how our technique simplifies the counterfactual information shown to users. The labels (b.1) to (b.7) refer to different information and functionality shown in the interface, see \ref{['sec-prototype']} for details.
• Figure 2: The defined casual graph in the synthetic data. The middle node is the target outcome, mortality risk, shown in red. The values near each link are the causal relationship strengths. The top 5 causal links are shown in red. Causal strengths range from 0.21 to 0.86, with a 0.05 interval between each causal strength in the graph.
• Figure 3: Four types of identified interaction behaviors in our exploratory analysis consist of atomic interactions. (a-b) are go-back behaviors and (c-d) are go-next behaviors.
• Figure 4: A search tree of a user in the CFACT group, with a width of 8, a filter-range layer of 8, a filter-variable layer width of 6, and a depth of 6.