Unraveling implicit human behavioral effects on dynamic characteristics of Covid-19 daily infection rates in Taiwan
share
We study Covid-19 spreading dynamics underlying 84 curves of daily Covid-19 infection rates pertaining to 84 districts belonging to the largest seven cities in Taiwan during her pristine surge period. Our computational developments begin with selecting and extracting 18 features from each smoothed district-specific curve. This step of computing effort allows unstructured data to be converted into structured data, with which we then demonstrate asymmetric growth and decline dynamics among all involved curves. Specifically, based on Theoretical Information measurements of conditional entropy and mutual information, we compute major factors of order-1 and order-2 that reveal significant effects on affecting the curves' peak value and curvature at peak, which are two essential features characterizing all the curves. Further, we investigate and demonstrate major factors determining the geographic and social-economic induced behavioral effects by encoding each of these 84 districts with two binary characteristics: North-vs-South and Unban-vs-suburban. Furthermore, based on this data-driven knowledge on the district scale, we go on to study fine-scale behavioral effects on infectious disease spreading through similarity among 96 age-group-specific curves of daily infection rate within 12 urban districts of Taipei and 12 suburban districts of New Taipei City, which counts for almost one-quarter of the island nation's total population. We conclude that human living, traveling, and working behaviors do implicitly affect the spreading dynamics of Covid-19 across Taiwan profoundly.
READ FULL TEXT
