Abstract
Social media provides a rich lens into how communities respond to disruptive economic and political events, but its unstructured scale creates challenges for conventional research approaches. To address this, Information Systems (IS) scholars have advanced Computationally Intensive Theory Construction, which combines domain knowledge, theory, and computational methods to surface theoretically meaningful patterns. This research-in-progress paper examines the 225 US tariffs using YouTube videos and comments collected between January and May 225. Emotions are identified and linked to discussion topics through emotion detection and topic modelling using gpt-4.1-mini and BERTopic, producing two proposed measures: the Emotion Index and the Emotion Polarity Index. These indices will be evaluated using ARX and VARX econometric models to assess their predictive power for asset returns. As work in progress, the study aims to demonstrate the feasibility of combining emotion–topic analysis with financial modelling, contributing to IS theory-building and offering new insights for Behavioural Finance.
Recommended Citation
Kok-Shun, Brice Valentin; Chan, Johnny; Peko, Gabrielle; and Sundaram, David, "Flaring Emotions, Hot Topics and Volatile Markets:
Modelling the Impacts of the 225 US Tariffs on Social
Media Perspectives and Asset Returns" (2025). ACIS 2025 Proceedings. 263.
https://aisel.aisnet.org/acis2025/263