| description abstract | This paper proposed and tested an integrated textual–visual–geo framework to enhance social sensing techniques in smart city digital twins in the context of disasters. Effective and efficient disaster response and recovery require reliable situational awareness regarding infrastructure disruptions and their societal impacts. Due to the rapid unfolding and evolution of events in disasters and emergencies, typical data sensing techniques (such as remote sensing and satellite images) are not sufficient to gain reliable situational awareness about disruptions that affect communities at a local scale. Social sensing enables gathering and analyzing massive user-generated data from various sources (social media, in particular) to monitor unfolding of localized events such as infrastructure disruptions and community needs. To advance social sensing methods and their integration into digital twins of cities, this study proposes an integrated framework for detecting infrastructure disruptions based on three information elements embedded in social media content: images, texts, and geo-maps. The framework consists of three main methods: a graph-based approach for detecting critical tweets, an image-ranking algorithm for selecting important images, and a kernel density estimate for estimating the geographical scales of the disruptions. The application of the proposed framework was demonstrated in a case study of water release from flood control reservoirs in Houston during Hurricane Harvey in 2017. The findings illustrate the capabilities of the proposed framework for capturing the critical situational information and interpreting the results for situational awareness and disruption response. The proposed framework can enhance integration of social sensing elements into smart city digital twins in the context of disasters. Accordingly, the proposed framework can improve the ability of community members, volunteer responders, residents, and other stakeholders in coping with built environment disruptions in disasters. | |
