Collaborative Research: Predicting Real-time Population Behavior during Hurricanes Synthesizing Data from Transportation Systems and Social Media

该项目开发了在自然灾害期间实时预测人口行为的新方法，有可能以具有成本效益的方式改变当前的应急响应状态。为了了解个人、基础设施系统和应急服务在此类灾害中应如何准备和响应，本项目利用了来自多个来源的数据，包括交通系统和在线社交媒体。使用创新的数据科学方法来整合来自多个来源的数据，可以提高应急响应预测和改进疏散交通管理可用数据的质量。将与从业人员分享研究成果，以促进紧急机构在飓风疏散和灾害管理方面改进决策。这一科研贡献支持了NSF促进科学进步和提高国家福利的使命。在这种情况下，收益将是改进应急响应的见解，这将挽救生命，减少经济损失，并减少未来事件中的恐慌，愤怒和混乱。该项目将来自交通系统和社交媒体的异构数据源整合到一个统一的框架中，为飓风期间人口动态行为的建模提供更好的信息。为了准确预测疏散需求，本项目利用了现有应急决策支持工具很少使用的大规模实时数据。它通过开发新的信息融合技术来表示人口及其行为，同时利用政府调查和社交媒体数据，文本挖掘方法从社交媒体数据中提取疏散意图，以及疏散交通预测模型来优化交通资源，从而推进灾害管理的数据科学。通过其创新的数据收集和建模方法，该项目将提高我们应对未来飓风的能力。该项目促使研究生和本科生更广泛地参与，包括来自代表性不足群体的学生，并计划向当地县市的交通工程师和应急管理官员更广泛地传播结果。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This project develops new methods to forecast real-time population behavior during natural disasters, potentially transforming the current state of emergency response in a cost-effective way. To understand how individuals, infrastructure systems, and emergency services should prepare and respond during such disasters, this project utilizes data available from multiple sources including from transportation systems and online social media. Using innovative data science approaches to integrate data from multiple sources increases the quality of the data available for emergency response prediction and improved evacuation traffic management. Research outputs will be shared with the practitioner community to facilitate improved decision making for emergency agencies in hurricane evacuation and disaster management. This scientific research contribution thus supports NSF's mission to promote the progress of science and to advance our national welfare. In this case, the benefits will be insights to improve emergency response, which will save lives, economic losses, and reduce panic, anger and confusion during a future event.The project combines heterogeneous data sources from transportation systems and social media, in a unified framework-providing better information for modeling dynamic population behavior during hurricanes. To accurately predict evacuation demand, this project leverages large-scale real-time data, rarely used by existing emergency decision support tools. It advances the data science of disaster management by developing novel information fusion techniques to represent population and its behavior while employing government survey and social media data, text-mining approaches to extract evacuation intent from social media data, and evacuation traffic prediction models to optimize transportation resources. Through its innovative data gathering and modeling approaches, this project will enhance our ability to deal with future hurricanes. The project engages a broader participation of graduate and undergraduate students including from under-represented groups and plans a broader dissemination of results to traffic engineers and emergency management officials from local counties and cities.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.