RAPID/Collaborative Research: Quantifying Social Media Data for Improved Modeling of Mitigation Strategies for the COVID-19 Pandemic

这笔快速反应研究(RAPID)赠款将支持研究，这些研究将为新冠肺炎大流行期间的社会行为和社区活动建模以及未来具有新冠肺炎特点的大流行贡献新的知识。该模式侧重于遵守缓解战略和公共卫生准则，从而能够选择在促进科学进步和促进国家健康与繁荣方面最有效的政策。目前正在使用各种大流行模型来预测病毒的传播，并确定哪种缓解策略是最有效的。这些模型在很大程度上建立在假设的基础上，可能包含了关于人口反应和行为的过于简单化的现实。这项研究将通过挖掘和分析新冠肺炎危机期间的社交媒体数据，为开发一个模型提供所需的知识和方法，该模型通过挖掘和分析社交媒体数据来研究美国个人对某些缓解策略的反应，这些策略包括社交距离、在家待命、隔离和旅行建议。这种改进的建模方法及其结果模型将对灾害应对管理人员和政策/决策者理解人类社会行为具有重要价值。这项工作可以评估流行病(和其他危机)期间缓解战略和公共卫生指导方针的有效性。该项目还将形成一个公开可用的案例研究的基础，该案例研究适用于大学水平的学生，可以广泛纳入课程。尽管基于个体和同类的混合流行病模型在一系列可能性中提供了有用的洞察力和预测能力，但它们对人们的行动高度敏感。这项研究旨在提供一种改进的方法来模拟大流行期间的社会行为和社区活动，以遵守缓解战略和公共卫生指南。社交媒体数据为研究公众如何以及在多大程度上遵守或不遵守建议的缓解战略和公共卫生指南提供了一个短暂的机会之窗。研究小组将从社交媒体收集与美国接触COVID19的地区人口相关的实时数据。这些数据将使用机器学习技术进行分析，以基于他们的人口、地理和时间信息的相似性来识别非互斥的人群，并在人群之间建立关系。分析后的数据将构成数据驱动的多范例模拟模型的基础，该模型捕捉公众情绪随时间的变化，量化对缓解战略和健康指南的抵制/遵守，并随着时间的推移衡量各种缓解战略和建议的整体有效性。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This Rapid Response Research (RAPID) grant will support research that will contribute new knowledge related to modeling social behavior and community activity during the COVID-19 pandemic, as well as future pandemics with COVID-19 characteristics. The model focuses on compliance with mitigation strategies and public health guidelines, thus enabling the selection of policies that are most effective in promoting both the progress of science and advancing national health and prosperity. Various pandemic models are currently being used to predict the spread of a virus and establish which mitigation strategies are the most effective. These models are heavily based on assumptions and may include an oversimplified reality of how populations react and behave. This research will provide needed knowledge and methods for the development of a model of how individuals in the U.S. react to certain mitigation strategies, such as social-distancing, stay-at-home orders, quarantines, and travel advisories, by mining and analyzing social media data during the COVID-19 crisis. This enhanced modeling approach and its resultant model will be of great value to disaster response managers and policy/decision makers to understand human social behavior. This work allows assessment of the effectiveness of mitigation strategies and public health guidelines during pandemics (and other crises). This project will also form the basis of a publicly available case study suitable for university level students that can be widely incorporated in courses. Although individual-based and homogeneous mixing pandemic models provide useful insights and predictive capabilities within a range of possibilities, they are highly sensitive to people’s actions. This research aims to provide an enhanced approach to model social behavior and community activity during a pandemic in terms of compliance with mitigation strategies and public health guidelines. Social media data present a brief window of opportunity for research on how, and to what extent, the public does or does not comply with the recommended mitigation strategies and public health guidelines. The research team will collect real-time data from social media related to COVID19-exposed regional populations in the U.S. The data will be analyzed using machine learning techniques to identify non-mutually exclusive clusters of people based on similarity of their demographic, geographic, and time information, and establish relationships among clusters. The analyzed data will form the basis of a data-driven multi-paradigm simulation model that captures changes in public sentiment over time, quantifies the resistance/compliance with mitigation strategies and health guidelines, and gauges overall effectiveness of various mitigation strategies and advice over time.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.