From puzzles to portraits: Enhancing situation awareness during natural disasters using a design science approach
基于情境意识理论，开发了一套处理海量社交媒体消息的协议，通过增量层次聚类和增强文本摘要算法生成简明摘要，帮助应急响应者快速掌握灾情全貌。
Abstract During emergency situations, a quick and concise summary from the deluge of messages improves “situation awareness” and enables informed decisions. However, this is challenging due to the volume, variety, and veracity of information. Grounded in the Situation Awareness Theory, this study presents a streamlined protocol to process millions of social media messages and render them in an easily comprehensible format for various stakeholders to gain deeper insights. Specifically, using a design science approach, we develop a set of artifacts using incremental hierarchical clustering and enhanced text summarization algorithm to produce informative summaries under crisis situations. We implemented our protocol on 2.5 and 3 million tweets collected during the two major hurricanes in 2017 and 2018. The results show that our protocol can derive critical information not captured by Twitter's search tools and mainstream news media and significantly improve on accuracy and efficiency when compared to other contemporary tools.