Capturing asymmetry in COVID-19 counts using an improved skewness measure for time series data

Capturing asymmetry among time series is an important area of research as it provides a range of information regarding the behaviour and distribution of the underlying series, which in turn proves to be useful for prediction. Classically, this can be achieved by modeling the skewness of the underlying series, usually using the standard measure. We present here an improved measure of skewness for time series which are integrated by a certain order, which is easy to calculate and proves to be advantageous over the existing one. We complement our methodology by implementing it to represent the heavy asymmetry among the daily COVID-19 case counts of several countries. • Improved skewness measure proves to be better than the usual skewness measure for time series data; • This new measure is applied on COVID-19 daily counts to capture the asymmetry appropriately.