RKI misreported results for either date=2021-01-25 or date=2021-01-26

[mathiasbynens]

Look at the numbers for Bayern:

• ~384k total cumulative doses reported for date=2021-01-25 (pubDate=2021-01-26)
• but only ~365k reported for date=2021-01-26 (pubDate=2021-01-27)

One of these must be wrong. My best guess is that RKI misreported the results for date=2021-01-25, and fixed it when reporting the date=2021-01-26 numbers.

[mathiasbynens]

I've confirmed that the data matches the RKI's documents (so the issue is not with our conversion scripts).

pubDate: 2021-01-26T10/49/24 date: 25.01.21 Gesamt:

pubDate: 2021-01-27T10/22/38 date: 26.01.21 Gesamt:

[mathiasbynens]

FYI @evilpie

[mathiasbynens]

Closing this for now. If the date=2021-01-26 / pubDate=2021-01-27 numbers are wrong, perhaps the RKI will publish an amended version later today (in which case our data will be corrected automatically). If not, then I guess the date=2021-01-25 / pubDate=2021-01-26 numbers were wrong instead.

[mathiasbynens]

Aha! The RKI has published an updated version of today’s spreadsheet, clarifying what’s going on:

* Enthält eine Korrektur der am 26.01.2021 veröffentlichten Daten für Bayern. Daher liegen die Werte heute niedriger als am 26.01.2021 berichtet.

So as we suspected, today’s numbers are accurate, and the date=2021-01-25 / pubDate=2021-01-26 numbers were inaccurate.