Possibly wrong recalculation per 100t in a group

[richtluk]

Describe the bug
It seems like there might be wrong recalculation of hospitalized cases per 100t group for vaccinated in chart on this page: https://ockovani.opendatalab.cz/statistiky_srovnani#hospitalizace

Could you please check that?

To Reproduce
Steps to reproduce the behavior:

1. Go to 'https://ockovani.opendatalab.cz/statistiky_srovnani#hospitalizace'
2. Click on 'Hospitalizovaní'
3. See smaller numbers for both vaccinated groups than unvaccinated although ÚZIS data https://twitter.com/dostalondrej/status/1554070813879541760 says otherwise.

Expected behavior
According to ÚZIS data (mentioned above) the numbers for vaccinated should be higher.

Screenshots
https://twitter.com/6m2xfrh28gisyq9/status/1554242198509592581

Desktop (please complete the following information):

• irrelevant

Smartphone (please complete the following information):

• irrelevant

Additional context
I haven't checked the code but it seems suspicious.

https://twitter.com/lukrichter3/status/1554535860250546176

[msusicky]

Thanks for this report, I will let you know soon why it is so.

[richtluk]

Hi, thanks for checking this out!

I had a quick look at the code where the calculation happens, but I didn't find any error there.

But clearly these are two different metrics. "Number of hospitalized patients" vs "number of newly hospitalized patients" and it is therefore possible that these behave differently depending on the vaccination.

[msusicky]

Take it as a hint, but we use: COVID-19: Týdenní přehled osob s prokázanou nákazou s ohledem na vykázaná očkování podle věkových kategorií from UZIS and I will bet a beer that it contains only people without previous infection.
Your source: not vacc. 3784 vacc. 12700 -> without previous infection and our data: 3701 x 12646. (1 day shift). I will ask UZIS and if it will be confirmed we will update the description.

[richtluk]

Hi, has UZIS already replied to you? I think your hypothesis makes sense. This might be it.

[msusicky]

Unfortunately yes, right on the 4th Aug. It is exactly like I wrote and they've said that new datasets will be ready with reinfections...............................................................
No, they are still not available :(

[richtluk]

Ouch. Thank you for the update!

So what now? Are you willing to write to them again? Or just update the description?