The rates are increasing because the cases/deaths are increasing as of the latest reported date you have. Now it's starting to decrease so your graph will show a decrease once you update it with the latest week data.
At least compare with a previous wave to see if there's a change according to vaccination rates (spoiler: there is), because as it is it's a worthless, albeit pretty, representation of data.
It is hard to make the comparisons I would like to make mostly because of the piss poor way the data are disseminated (or not disseminated).
For example, the CDC overwrites the previous information with every week's update. That is why the repo exists. To preserve any time series information in one place with a verifiable way to extract it out of ever-changing HTML pages and put it in a table.
In theory, the CDC ought to be able to put the relevant information in a table. At a minimum, we'd need `date`, `locality`, age distributions of fully, partially, and never vaccinated populations, age distribution of people who are hospitalized due to COVID19, and age distribution of people who died from COVID19.
Ideally, we'd have a data set consisting of one row per hospitalization/death with relevant dimensions such as age, sex, first shot date, second shot date, type of vaccine, locality, other conditions etc.
But, we do not. Because the bureaucracy has not deemed it appropriate to collect or share that information.
Instead, we need to rely on free form HTML pages where the provenance of the data are not clear.
What I have done is made visible one bit of information that would otherwise not be available: The normalized rate of fully vaccinated people being hospitalized or dying from the disease against which they are fully vaccinated has been steadily increasing over time.
The main reason is that the vaccines are not as effective as the 95% number that has been flouted time and again. Of course, it takes time for a person to be exposed, diagnosed, hospitalized, and maybe die, so the revelation has been happening over time instead of instantaneously.
> because as it is it's a worthless, albeit pretty, representation of data.
If I hadn't taken the snapshots and extracted the information from those pages, there would be no time series of breakthrough hospitalization and death rates anywhere. It seems to me it is worth something to save that information.
Plus, Biden told us the rate of hospitalization from COVID19 among the fully vaccinated was 1/160,000. The real number is 10 times that[1]. Isn't it worth something to know this?
At least compare with a previous wave to see if there's a change according to vaccination rates (spoiler: there is), because as it is it's a worthless, albeit pretty, representation of data.