February 20th, 2021


Interesting Links for 20-02-2021


Rate of change of Scottish infections

Warning up-front: I am not a statistician. I'm just playing with the numbers here to scratch an itch, and I'm happy to be told that there are better ways of looking at this.

However, I was interested to see what the change over time was for the case count in Scotland.

So I grabbed the data from the Scottish government publication page, extracted the 7-day case count for each day, and then compared that to the 7-day case count from a week earlier, and stuck it in a graph.

The 7-day case count is useful because there are all sorts of artefacts that occur on a daily basis, including it taking longer for cases at the weekend to be reported. (I check here on a daily basis around 15:00, and you can see the daily figures vary in a similar pattern each week.) And comparing it with a week earlier means you get a good idea of whether there's a general upward or downward trend - whereas daily trends can be a bit deceiving.

The really big jumps up are mostly due to very low figures. The 211% jump was from 14 cases to 30 - a minor outbreak occurring. The scary one is in December, when the cases almost doubled from 1,200 to 2,300.

From there you can see a steep drop at the start of this year. The current worrying bit isn't a peak at all. It's that the 25% weekly drops we'd been seeing through January and early February have now gone fairly flat. Here's the 1-week changes for each day in the last week. Whatever we were doing, it's not working as well as it was...

Date1-week change

Original post on Dreamwidth - there are comment count unavailable comments there.