Social Media Communication Patterns of Construction Industry in Major Disasters

Construction industry stakeholders tend to adopt innovative strategies for the smooth operation of their project delivery mechanisms. However, during major disasters such as hurricanes, the construction industry experiences several challenges in responding to an approaching hurricane as well as the aftermath recovery. Online social media platforms are becoming more prevalent for effective information dissemination, and recent studies suggest that social media plays an important role in crisis communication. The primary objective of this study is to reveal social media communication patterns of construction industry during Hurricane Michael (2018). By employing Twitter application programming interface (API), similar to 275 Million tweets (from Sep. 9 to Nov. 19, 2018) were collected using relevant, keywords such as hurricane, construction, workers, safety, site, subcontractor, hazard, rebuilding, schedule, supply, among others. Several machine learning techniques (topic model and dynamic topics over time) and data mining approaches (most frequent words) have been used to analyze user sentiments and concerns over time and space based on text, geo-locations, and timestamps of such tweets. While the hurricane paths were critical but uncertain to follow over time, this data-driven method revealed crisis communication patterns of construction industry. Specifically, the necessity of planned construction during hurricane, construction at night before landfall, continuous job opportunity in construction industry, and project closure after landfall are among prominent topics. The crisis communication patterns obtained in this study will help construction stakeholders and policymakers to make strategic decisions that are community-centered as well as to facilitate effective response and recovery in major disasters. This type of study can also reveal latent challenges that delay reconstruction of infrastructure and provide opportunities to improve recovery time.