Since 2020 is a very morbid year, let’s look at monthly absolute deaths for Sweden, years 2000 – 2019:
Look’s pretty messy. But with a bit of imagination, we might start to suspect there might be some pattern, some periodicity in the data.
So let’s take a very Big Hammer, probably the most important tool of maths ever, the Fourier Transform, to help us detect any possible patterns in the data:
Sure enough : there *is* a pattern (actually several, but let’s focus on the most significant one) : a yearly pattern of deaths as shown by the frequency domain chart above. That is, over a year, deaths tend to rise and fall, that is, seasonality is very much present in Swedish death data.
Don’t believe me…? Below a zoom-in on 2015-2019, which might make it easier to see the yearly periodicity.