@seb No thats not how data analysis works. You do not count the number of people That would be a dishonest and inaccurate representation. It is what we would call non-normalized data in the trade (remember this is what I do for a living).
As for looking at it against the population percentage, again, thats not an honest way to present the data either when we talk about cause and effect tests (things like the Granger Causality test). since we are trying to measure the causality between a change in policy (ending lockdown quickly) and its effect on the infection rate then the honest and accurate way to look at that is the rate of change of infection rate. It is the only way to make a case that suggests causality rather than simple correlation. Read up on the Granger Causality test to understand this idea a bit more if you'd like.