IEEE VIS 2024 Content: Visual Analysis of Multi-outcome Causal Graphs

Visual Analysis of Multi-outcome Causal Graphs



Mengjie Fan - Institute of Medical Technology, Peking University Health Science Center, Beijing, China. National Institute of Health Data Science, Peking University, Beijing, China

 Jinlu Yu - Chalmers University of Technology, Gothenburg, Sweden. Peking University, Beijing, China

 Daniel Weiskopf - University of Stuttgart, Stuttgart, Germany

 Nan Cao - Tongji College of Design and Innovation, Shanghai, China

 Huaiyu Wang - Beijing University of Chinese Medicine, Beijing, China

 Liang Zhou - Peking University, Beijing, China

 Screen-reader Accessible PDF

 Download Supplemental Material

 Room: Bayshore VII

2024-10-18T12:30:00ZGMT-0600Change your timezone on the schedule page
2024-10-18T12:30:00Z

Exemplar figure, described by caption below — The case study of the UK Biobank data with a medical expert using our method. In the first stage of "single causal graph analysis" (1–4), the expert explores and edits single causal graphs using the progressive comparative visualization of three state-of-the-art causal discovery techniques (2-4) in combination with her domain knowledge. In the second stage of "multi-outcome causal graphs comparison" (5, 6), she selects graphs of interested outcome for comparison using various layouts, including the supergraph (5), and our new comparable layout for subgraphs (6).

Fast forward

Full Video

Keywords

Causal graph visualization and visual analysis, causal discovery, comparative visualization, visual analysis in medicine

Abstract

We introduce a visual analysis method for multiple causal graphs with different outcome variables, namely, multi-outcome causal graphs. Multi-outcome causal graphs are important in healthcare for understanding multimorbidity and comorbidity. To support the visual analysis, we collaborated with medical experts to devise two comparative visualization techniques at different stages of the analysis process. First, a progressive visualization method is proposed for comparing multiple state-of-the-art causal discovery algorithms. The method can handle mixed-type datasets comprising both continuous and categorical variables and assist in the creation of a fine-tuned causal graph of a single outcome. Second, a comparative graph layout technique and specialized visual encodings are devised for the quick comparison of multiple causal graphs. In our visual analysis approach, analysts start by building individual causal graphs for each outcome variable, and then, multi-outcome causal graphs are generated and visualized with our comparative technique for analyzing differences and commonalities of these causal graphs. Evaluation includes quantitative measurements on benchmark datasets, a case study with a medical expert, and expert user studies with real-world health research data.