| description abstract | Social media, a near-real-time information source, has been widely used to provide timely infrastructure-related insights following disasters. However, most previous social media–based studies have been solely focused on assessing infrastructure damage, which leads to partial situational assessment lacking the other main types of infrastructure conditions, i.e., functioning and restoration. To bridge this research gap, the present study aims to examine the use of social media for systematically sensing infrastructure conditions following disasters. An efficient topic modeling–based approach is proposed (1) to model infrastructure condition–related topics through incorporating domain knowledge into the correlation explanation, and (2) to investigate the spatiotemporal patterns of topic engagement levels for systematically sensing infrastructure functioning, damage, and restoration conditions. To demonstrate the feasibility and applicability of the proposed approach, electricity infrastructure conditions within Florida following Hurricane Irma are studied. This research provides a systematic situational assessment of critical infrastructure following disasters, thereby enabling practitioners to make informed decisions in infrastructure management. Also, the proposed approach investigates the interactions between humans and infrastructure systems, which advances human-centered infrastructure management. | |
