SMG-Vis: Supporting Judgment in Multi-Target Social Media Analysis through Group Exploration and Stance Reasoning
Authors
Yihang Yang (Fudan University), Kengyi Wang (Fudan University), Chen Yi (Fudan University), Chunran Hu (Fudan University), Xingyu Lan (Fudan University), Siming Chen (Fudan University)
Presentation
- Session
- My followers need to know about this!
- Time
- Thursday, Nov 12, 10:48 – 11:00 (US/Eastern) · session 10:00 – 11:30
- Location
- Hall America center
Keywords
Social media, visual analytics, stance analysis, multi-target stance analysis
Abstract
Social media analysis is widely used by domain experts to understand public stances on complex issues. However, analyses that focus on isolated targets or single-perspective groupings can lead to biased interpretations and misleading attributions. We present SMG-Vis, an interactive visual analytics system that supports expert judgment in multi-target social media analysis. SMG-Vis represents each user as a stance signature over multiple targets and constructs a multi-target stance space in which groups are defined by identical stance signatures. This representation enables analysts to explore structured group relationships formed by combinations of stances across targets. SMG-Vis supports iterative group exploration through coordinated views of mentioned targets, keywords, and supporting posts. To assist reasoning, the system provides computationally assisted cues that expose alternative explanations and enable comparison of competing hypotheses across groups. By juxtaposing focal groups with related or contrasting groups, SMG-Vis helps analysts assess the robustness of their interpretations and mitigate misinterpretations caused by limited analytical perspectives. We demonstrate the utility of SMG-Vis through real-world case studies on social media datasets covering major public issues, showing how it supports nuanced group understanding and more reliable analytical judgments.
For Practitioners
This paper is intended for practitioners who analyze complex social media discussions, including policy analysts, communication researchers, political scientists, and open-source intelligence (OSINT) analysts. They can apply the proposed workflow to explore multi-target stance spaces, identify stance-defined groups, inspect supporting evidence, and evaluate alternative explanations to support transparent and evidence-based judgment.