The COVID-19 pandemic

[tomh009]

Excel only allows two vertical axes, and I have the number of cases on one and the two percentages on the other. If I user a linear scale for the percentages, it is quite difficult to see movement in the positivity rate as the % of actives is roughly 5x the positivity. Not ideal this way, but I decided it was better than the other way

Positivity is up in the last five days, but that's partly because the number of tests is down by 20-25%. The two are interlinked and it's hard to know how exactly much of the positivity to attribute to the lower number of tests.

Indeed, that is my question as well.

I wonder how much validity there is in simply scaling the cases by the positivity rate?

I think that's too simplistic. Generally the high-risk people (with exposure risk or symptoms) are most likely to get tested, so the fewer people you test, the higher the positivity rate will be. As you test more and more members of the general public (without any specific risks) you will get closer to a positivity rate that matches the percentage of the population that is infected (and not yet confirmed).

But I can't predict what that curve looks like. Maybe someone on the Internet has done this work but I have not yet stumbled across it.