Downloading Doom

Raiding the Office for National Statistics' fridge

This morning I have spent a fair few unshaven and somewhat unkempt hours grabbing weekly death data (all cause and COVID by age band and gender) for England & Wales from the Office for National Statistics. I wasn't best pleased to learn they're altered the format of their spreadsheets for 2022 but confess to liking the new format even better.

Prior to 2020 age banding came in seven lumps (U1, 1-14, 15-44, 45-64, 65-74, 75-84, 85+) and I've retained this coding to a permit historic analysis dating back to 2010. From 2020 onward the ONS shifted up a gear, offering analysts twenty lumps (U1, 1-4, 5-9... 85-89, 90+), so I've shovelled that lot into the bucket as well.

In that same bucket we have comorbid and causal tallies for respiratory death, influenza/pneumonia and COVID along with an indicator variable marking changes to ICD10 definitions for COVID (they extended the scope back on 27 Feb 2021). Total population estimates are in there as well but so far I’m avoiding those murky estimates by age band.

With a contiguous data series now spanning 640 weeks I'll be able to whip out some fancy time series techniques to see where that gets us, but we'll start out keeping things seriously simple.

Kettle on!