Using ARIMA To Investigate COVID Death (part 5)

Cranking the handle on the latest daily data sitting in the UK GOV coronavirus dashboard

We arrive at a potentially juicy piece of analysis, being ARIMA modelling of daily certified COVID death across England for the period 23 May 2021 - 12 Jul 2022. Let’s start with a plot of the dependent variable:

Now that is the strangest time series for deaths I have ever seen! In addition to the expected 2021/22 winter peak we’ve got a summer peak, autumn peak, spring peak, and what appears to be the beginning of a second summer peak. Assuming these deaths were appropriately certified by experienced physicians working according to training rather than management whimsy, and assuming emphasis hasn’t been placed on test results, then we may conclude that we’re looking at a strangely cyclic catastrophe.

I say ‘strange’ because some but not all of these peaks in daily deaths can attributed to a rise in COVID cases as the following dual time series plot reveals: