John D. Tickle Professor Joshua Fu is the latest CEE faculty member to focus his efforts on the COVID-19 pandemic, which has continued to spread in the US as other parts of the world have managed to contain the outbreak. While the US was surging to the top with most number of cases, there was still a push by several states to re-open business.
Fu and his doctoral student Cheng-Pin Kuo were able to take available data to build a predictive model that highlights that if authorities are to issue lockdowns, they should last more than one week to be effective.
The research is the first attempted COVID-19 case predicting model based on county-level demographic, environmental, and mobility data constructed with multiple machine learning techniques and a hybrid framework. His data encompasses the insights gained can provide a quantitative reference for policy approaches with this and future pandemics.
It looked at all counties in the US as well as compared selected metropolitan counties including New York City, Cook County in Illinois, Los Angeles County in California, and Miami-Dade County in Florida to assess the impact from lockdown, Phase I, and Phase III re-opening.
In utilizing available data, Fu and Kuo noticed a trend that suggested that the community mobility decreased substantially after the lockdown but kept increasing with an apparent weekly pattern that implied high infections during the weekend.
The results indicated that, compared with Phase I re-opening, a one-week and a two-week lockdown could reduce infections between 4–29 percent and 15–55 percent, respectively. Meanwhile, a two-week Phase III re-opening could increase infections 16–80 percent. Based on the findings, Fu and Kuo concluded that the mandatory orders in metropolitan counties such lockdown should last longer than one week.
The impact of lockdown or re-opening was also county-dependent and varied with the local pandemic. In the future, Fu expects to have higher quality data with more data points and longer timeframes to evaluate.
He will also consider more county-dependent factors and employ more sophisticated techniques to decrease the modeling uncertainty and apply it to counties nationally and other countries.