CasesByAge
Short explanation
(These graphs should auto-update daily at around 4-5pm UK time. You may need to hit shift-reload or some such to defeat your browser's cache.)
30th December:
- Warning: Case numbers may be limited by testing capacity now. See https://russss.github.io/covidtracker/testing.html for some evidence of this.
29th December:
- Warning 1: data over Christmas period may be misleading owing to unusual testing and reporting patterns. There is special case processing here which moves cases around between 25th, 26th and 27th December (conserving the total number of cases in each age band). This is just my guess to produce something somewhat sensible by attempting to undo the effect of delayed tests over Christmas. It looks by eye to me like it removes a decent chunk of the (otherwise large) artefact over Christmas, but there is absolutely no guarantee this a correct procedure and there is no guarantee that there aren't other recent artefacts, such as change in overall propensity to test.
- Warning 2: The last one or two values of each graph are less certain than the others and it is risky to use these alone to gauge whether cases are going up or down.
The idea of this graph is to attempt to make use of the most recent information from public data about the number of confirmed cases in various age bands, so that trends over the last few days can hopefully be discerned. The numbers in each age band are adjusted/corrected for their incompleteness, then corrected for day-of-week biases, in both cases, as far as possible, in a non-laggy way.
The value of this approach is that it can show trends ahead of many other indicators, and in a reasonably clear way without the day-of-week effects muddling it up. On the other hand, the possible weakness is that it relies on case counts being a reliable indicator of the number of infections, which would cease to be true if people changed their willingness to get tested or our testing system were overloaded. As of 22 December 2021, it's my opinion that the testing system is showing signs of strain but is broadly speaking holding up. This proposition will eventually be checked one way or another by infection survey data.
Note that, as a consequence of the last few days' information being incomplete, the case number estimate for these days can sometimes be retrospectively updated over the next day or two.
Program: https://github.com/alex1770/Covid-19/blob/master/casesbyage.py
England
Regions of England
