Project 2
Project 2 examines the effect of movement-based restrictions on mobility in Nigeria and South Africa, comparing pre-pandemic and post-pandemic patterns in cell phone mobility data and connecting specific movement patterns with COVID-19 risk. This model incorporates state-of-the-art mobility analytics from the transportation sector, applying them to the African context, possibly for the first time.
PRINCIPAL INVESTIGATOR
Chenfeng Xiong, PhD
Chenfeng Xiong is an assistant Professor at the Department of Civil and Environmental Engineering, Villanova University. He holds a Bachelor’s degree in Civil Engineering from Tsinghua University, and an M.S degree in Civil Engineering, an M.A in Economics and a Ph.D in Transportation Engineering, from the University of Maryland. Chenfeng’s areas of interest include, transportation big data, human mobility and travel behavior modeling, transportation economics, carbon neutralization, large-scale agent-based analysis modeling and simulation (AMS). As a researcher, Chenfeng has published over 80 peer reviewed articles.
Dr. Xiong also leads one of the INFORM Africa pilot projects and currently serves as the Chairperson for the DSI-Africa steering committee.
TEAM MEMBERS
ACTIVITIES
The Project 2 Research Team employed mobile device location data and developed in-house big-data and data-driven algorithms to analyse human mobility, activities, and population density in South Africa and Nigeria during the SARS-CoV-2 pandemic. The team’s cloud-based big-data computing infrastructure ingests terabytes of data daily and produces those high-resolution human mobility measurements.
Density Plot of 14 Days of Human Activities in South Africa
The mobility measurements produced by the team are new and have filled a major data gap in understanding the travel behaviour change during the SARS-CoV-2 in Nigeria and South Africa. The team assessed the effectiveness of mobility restricting policies as key lessons learned from the pandemic and found that travel bans, and federal lockdown policies failed to restrict trip-making behaviour but had a significant impact on distance travelled.
Daily trips per person and daily distance travelled per person measured using smartphone location data collected in Nigeria (January 01, 2020 – April 25, 2020).
Using the developed analytical framework in the previous project years, the team has produced individual-level trip rosters. The roster information was used to calculate the metrics of three types of human mobilities (shown in Figure 1), including provincial inflows, cross-district flows, and within-district flows, for all provinces in South Africa. These metrics were used to evaluate the dynamic impacts of three types of human mobilities on daily reported COVID-19 cases for 2020.
Figure 1-(a): illustration of 3 types of human mobility flow
Figure 1-(b): # of Type 1 trips. Before & After COVID-19 comparison
Figure 1-(c): # of Type 2 trips. Before & After COVID-19 comparison
Figure 1-(d): # of Type 3 trips. Before & After COVID-19 comparison
The team built a structural equation model at province-level of South Africa to analyze the impact of three types of mobility metrics on COVID-19. The varying patterns of the metrics are shown in Figure 2.
Figure 2: (a) Varying pattern of the three types of mobility metrics of South Africa (weighted average by province population) and daily reported new COVID-19 cases from March 5th 2020 to December 31st 2020. (b) Separated by provinces of South Africa.
Through our dynamic structural equation model, we observed that at the emergent stage of the pandemic, the government policy of South Africa played an important role in guiding people’s travel behavior. However, on May 1st, the lockdown level was lowered from level-5 to level-4, we observed that the impact of the level-4 lockdown policy on people’s travel behavior became extremely limited, It shows that the acceptance of the travel restriction policy was substantially decreased among the people. We believe it can be attributed to the reaction of people’s fatigue to the social distancing as people have been restricted at home for a long time.
Figure 3: Random effects of lockdown level on natural-logged three types of mobility metrics
The team has built a structural equation model at province-level of South Africa to analyze the impact of three types of mobility metrics on COVID-19. The team incorporated origin-destination information into the three types of mobility metrics and captured their time-varying impact on the daily new cases. Through using the three types of human mobilities demonstrated in Aim 1, the dynamic relationships are shown as follows in Figure 4.
