Tweeting Through the Flood: Application of BERT Topic Modeling for a Comparative Flood Communication Analysis

Authors

Keywords:

Crisis informatics, flooding, natural language processing (NLP), social media, Twitter/X

Abstract

Floods are prevalent disasters in the United States, posing escalating risks due to climate change-induced factors like rising sea levels and erratic rainfall patterns. Despite governmental efforts, flood risk communication remains inadequate, hindering preparedness and response capacity. While governmental agencies predominantly employ traditional, one-sided information dissemination approaches, social media platforms pose as crucial early warning indicators and sources of real-time information. This study conducts the first part of our intended comparative analysis of social media messages between governmental agencies and communities. We perform a case study on a flooding event in Michigan from May 17-20, 2020. Utilizing advanced topic modeling, we examine Twitter/X message content and sentiment from community-based posts. Insights aim to inform more effective flood communication strategies, bridging the gap between official information and community needs during events. Future work will compare social media messages' content and sentiment from governmental agencies relative to those from the community.

Downloads

Download data is not yet available.

Downloads

Published

2024-06-26

How to Cite

Salley, C., Fox, N. ., & Schubert, A. (2024). Tweeting Through the Flood: Application of BERT Topic Modeling for a Comparative Flood Communication Analysis. ISCRAM Proceedings, 21. http://ojs.iscram.org/index.php/Proceedings/article/view/115

Similar Articles

1-10 of 50

You may also start an advanced similarity search for this article.

Most read articles by the same author(s)

1 2 3 4 5 6 7 8 9 10 > >>