Spatiotemporal Insights into Online Public Responses to Disasters in Developed and Underdeveloped Countries

Early-warnings and situational awareness during disasters are critical for protecting our built environments and communities. In recent years, the increasing prevalence of social media platforms has presented an unprecedented opportunity for gathering real-time information, facilitating rapid hazard detection, evacuation plan propagation, and damage assessment and recovery. However, underdeveloped countries have been inadequately explored compared with developed countries across different spatiotemporal scales. To address this knowledge gap, this study investigated online public responses before, during, and after two recent hurricanes, aiming to identify the different patterns between underdeveloped and developed countries at different stages of disasters. Twitter data of Hurricane Eta in Honduras and Nicaragua and Hurricane Ian in the US was collected and analyzed using sentiment analysis and topic modeling. The findings revealed that individuals in Honduras and Nicaragua tended to express more negative attitudes and had more concerns about disaster damage than those in the US. Sentiment differences were also found between moderately and mostly affected regions in the US. Recommendations for future research directions and disaster response practices were proposed based on the findings.