Advancing Methods to Trace and Contextualize Space-Time Interaction Patterns in Movement Data

该研究项目将推进计算方法，以跟踪和表征移动网络中代理之间的交互和关键遭遇。网络的例子包括城市中的人、生态系统中的一群动物或船队。尽管跟踪技术取得了进步，但计算运动分析方法在大型移动网络中动态交互模式的量化和表征方面仍然有限。进入21世纪20年代，社会目睹了SARS-CoV-2通过密切接触和/或个体之间的滞后互动通过呼吸道飞沫广泛传播。这导致了一系列前所未有的非药物干预措施，包括数字接触者追踪，以减轻COVID-19的传播。然而，目前的技术在追踪和检测健康和潜在感染个体之间的关键或危险接触或暂时滞后的相互作用方面效率低下。使用运动观察，该项目将提供关于移动代理之间相互作用的数据驱动结果。研究结果将加强追踪接触者的技术，以检查人类可能接触的健康风险或传染媒介。更一般地说，即将开发的方法将使科学家能够模拟人类和动物网络中的社会行为。该项目将创建开放获取/开源分析工具，使地理和其他领域的研究人员、教育工作者和学生更容易获得空间数据科学。该项目将为研究生提供培训和研究经验。本研究将通过优化的计算算法开发和评估新的环境感知时间地理分析方法，以(1)追踪动态交互，并使用大型运动数据集测量个体之间相遇的持续时间和频率；(2)将相遇、并发交互和滞后交互置于环境中，以更好地识别关键或危险的接触。本研究将探讨三个主要的研究问题：(1)我们如何最好地利用统计方法和时间地理方法来更好地估计通过运动产生的接触？(2)考虑到大规模的运动观察，我们如何有效地追踪和识别个体之间的“风险”或“有趣”相遇？(3)交互分析能否用于理解人类和动物社会网络中的集体运动模式？将开发一套案例研究和开放分析工具，以利用对人和动物的真实GPS观测来证明分析框架的有效性。本研究发展的分析方法将推广到理解社会和生态系统的相互作用，为人类社会行为和关键物种竞争提供新的知识。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 research project will advance computational approaches to trace and characterize interactions and critical encounters between agents in a mobile network. Examples of networks include people in a city, a group of animals in an ecosystem, or a fleet of vessels. Despite the advances in tracking technologies, computational movement analysis methods remain limited in quantification and characterization of dynamic interaction patterns in large mobile networks. As the decade turned to the 2020s, society witnessed the widespread transmission of SARS-CoV-2 through respiratory droplets via close contacts and or lagged interactions between individuals. This led to a set of unprecedented non-pharmaceutical interventions including digital contact tracing to mitigate the spread of the COVID-19. However, current techniques are inefficient for tracing and detecting critical or risky encounters or temporally lagged interactions between healthy and potentially infected individuals. Using movement observations, this project will provide data-driven results about interactions between moving agents. The results will enhance contact-tracing technologies for examining potential human exposure to health risks or infectious agents. More generally, the methods to be developed will enable scientists to model social behaviors in human and animal networks. The project will create open-access/open-source analytical tools which will make spatial data science more accessible to researchers, educators, and students in geography and other fields. The project will provide training and research experiences for graduate students.This research will develop and evaluate novel context-aware time-geographic analytical methods through optimized computational algorithms to (1) trace dynamic interactions and measure the duration and frequency of encounters between individuals using large movement data sets, and (2) to contextualize encounters, concurrent interactions, and lagged interactions to better identify critical or risky contacts. The research will investigate three overarching research questions: (1) How can we best leverage statistical approaches and time geographic methods for better estimation of contact through movement? (2) Given large movement observations, how can we effectively and efficiently trace and identify 'risky' or 'interesting' encounters between individuals? (3) Can interaction analytics be used to understand collective movement patterns in social networks of humans and animals? A set of case studies and open analytical tools will be developed to demonstrate the efficacy of the analytical framework using real GPS observations of people and animals. The analytical methods to be developed in this study will be generalizable to understanding interaction in both social and ecological systems, contributing new knowledge about social behavior of humans and competition of keystone species.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.