Doctor of Philosophy (PhD)
Disaster resilience is the capacity of a community to ‘bounce back’ from disastrous events, by effecting resilience building activities through the four phases of emergency management – preparedness, response, recovery, and mitigation. However, data describing a community’s behaviors in each phase are difficult to access in traditional databases. Social media data provide an innovative data source to better understand communities’ responses and behaviors in each phase of emergency management during disasters, and thus could ultimately monitor and enhance disaster resilience. This study analyzes the spatial-temporal patterns of Twitter activities during Hurricane Sandy, which struck the northeastern United States on October 29, 2012. The objectives of this dissertation research are to: (1) develop a methodological framework to extract a set of common indices from Twitter data, a type of Big Data, for emergency management and resilience analysis; (2) examine if there are significant geographical and social disparities in Twitter use through the three main phases of emergency management; (3) test if social media data could help estimate post-disaster damage; and (4) test if Twitter indices can predict disaster resilience. Four corresponding hypotheses were tested. Results show that common indices derived from Twitter data, including Adjusted Frequency, Ratio, Sentiment, and Normalized Ratio, could enable comparison across regions and events and should be documented. Social and geographical disparities in Twitter use were found to exist in the Hurricane Sandy event. Communities closer to the landfall location and coastline, which suffered higher levels of threat from the disaster, show higher Twitter activities, especially in the preparedness phase. Under the same level of threat, communities with better socioeconomic status show higher Twitter activities, particularly in the response phase. Under the same level of threat, communities suffered more damage tend to have higher Twitter activities and more negative sentiment towards Hurricane Sandy in the recovery phase. Adding Twitter indices into a damage estimation model improved the adjusted R2 from 0.46 to 0.56, indicating that social media data could help improve the post-disaster damage estimation. Disaster resilience assessment using the Resilience Inference Measurement (RIM) model shows that counties with higher levels of community resilience (usurper and resistant) formed clusters in areas near New York city. Susceptible counties were mostly in New Jersey, which is adjacent to the hurricane landfall location. The resilience score was positively correlated with “mean value of owner occupied housing units”, “percent of housing units built after year 2000”, and “percent of population over 25-year old with a bachelor or higher degree”, whereas “percent of mobile homes”, “percent of civilian labor forces that are not employed”, and “median age” were negatively correlated with the resilience score. The correlations between Ratio indices and resilience scores were significant in the response and recovery phases, which confirms that Twitter data has the potential to predict disaster resilience. The methodology developed could be used to compare different disastrous events, and knowledge gained from this study could provide valuable insights into strategies of utilizing social media data to reduce disparities and build long-term resilience to disasters.
Document Availability at the Time of Submission
Secure the entire work for patent and/or proprietary purposes for a period of one year. Student has submitted appropriate documentation which states: During this period the copyright owner also agrees not to exercise her/his ownership rights, including public use in works, without prior authorization from LSU. At the end of the one year period, either we or LSU may request an automatic extension for one additional year. At the end of the one year secure period (or its extension, if such is requested), the work will be released for access worldwide.
Zou, Lei, "Mining Social Media Data for Improved Understanding of Disaster Resilience" (2017). LSU Doctoral Dissertations. 4274.
Lam, Nina S
Available for download on Saturday, February 23, 2019