Examining Crowdsourced Data for Improving Flood Emergency Management

Abstract

Natural hazards, such as flooding and hurricanes, wildfires, pose a threat to the well-being of communities by disrupting the infrastructure services people use in their day-to-day lives. To mitigate these impacts, Big Data has been used extensively in research to improve the response and recovery during the aftermath of a disaster and to improve the performance of infrastructure systems. This has provided a lot of promise and innovation to the research field of flood emergency management. However, current approaches have done little to assess how to effectively integrate Big Data into infrastructure systems and look at the multiple biases that inherently come with Big Data. To overcome these challenges, I developed analytical frameworks, provided empirical models, and utilized spatial analysis techniques. In the first study, I proposed and tested a framework to examine the extent that crowdsourced data can improve flood monitoring by assessing if the data can supplement flood gauges to monitor transportation infrastructure systems. The second study develops regression models to address which socially vulnerable populations are not represented on crowdsourced data platforms, spatial maps to identify sample and spatial data imbalances and assess which socially vulnerable groups are impacted by these data imbalances. The created frameworks, findings, methodologies, and models in this research have significant contributions for improving flood emergency management by enabling integration of crowdsourced data into infrastructure systems and addresses the biases that come with the data to create a holistic and equitable recovery plan.