Purdue University

    Data Science, Human Behavior and Improving Responses to COVID-19

    May 29, 2020

    COVID-19 loves densely populated spaces.  As we have seen in situations such as Mardi Gras celebrations in New Orleans, spring break beach parties in Miami, Florida, and even large funerals in Williamsburg, New York, peoples’ behaviors, both individually and collectively, provide keys to how governments and policy makers might need to suggest, regulate or even mandate public movement and behavior to mitigate virus transmission.  Two current studies involving faculty affiliates are investigating this through analysis of mobile phone data and through the creation of a software infrastructure that will help scientists investigate the risk of the spread of COVID-19 and suggest how we may better prepare for future epidemics in light of high-density locations.

    Satish Ukkusuri, Professor of Civil Engineering, has been involved in a team project with four researchers from Japan, including one from Yahoo! Corporation.  In a recently pre-published preliminary study, the team used large-scale mobility data collected from more than 200K mobile phones in Tokyo, Japan in the one week after the Japanese government declared a state of emergency (ending April 15th).  This data gave the team an opportunity to monitor and understand the impacts on human mobility given non-compulsory recommendations to reduce movement and ‘shelter-in-place’.  Their preliminary findings, published May 18th, suggest that these measures reduced mobility by over 50% and that they had an impact on subsequent decrease of the effective reproductive number of COVID-19. While their results are limited in scope, for example, they could not consider factors such as increased hand washing and other sanitizing measures, and could not isolate correlated factors such as potential age bias in users themselves, the study provides an interested basis for future work by comparing results with other datasets from other cities in Japan or elsewhere.

    Faculty affiliate David Ebert, Silicon Valley Professor of Electrical and Computer Engineering, is working with a team of researchers led by Yung-Hsiang Lu (ECE) as well as David Barbarash (HORT) and Wei Zakharov (Library Science) on an NSF-funded RAPID grant entitled “Leveraging New Data Sources to Analyze the Risk of COVID-19 in Crowded Locations.”  For this grant project, the team is looking at possibilities for maintaining social distance through fine scaled analysis of the movement and interaction patterns of people at crowded locations that may suggest interventions, such as changes to crowd management procedures and the design of built environments, that can yield social distance without being as disruptive to human activities and the economy as current complete shutdowns have been. Their work will provide human data needed to further develop mathematical models on pedestrian dynamics.  The team will collect data from available data streams, such as public webcams and location- based services. The project team will also perform outreach to decision makers so that the scientific insights yield actionable policies contributing to public health.

    Related to the need for data about peoples’ movements and behavior, we need data about individuals themselves such as what is gained through testing for the virus and managing results in an effort to improve national and global efforts to improve the logistics and optimize testing.  Center affiliate Chad Laux, associate professor of Computer and Information Technology at Purdue, has been putting his expertise in Lean Six Sigma, a method that relies on a collaborative team effort to improve performance by systematically removing waste and reducing variation, to think about testing protocols for COVID-19.  With collaborators from around the world, Laux is working with his team to think about how big data solutions using Lean Six Sigma principles can help improve data analysis and implement machine learning and AI approaches to improve testing and tracking and thus improve flexible responses with fewer false positives and less time wasted.

    Author: Lynne Dahmen

    Back to news