WARNING for all #covid19sverige followers! This is likely to be a problem for the next 1-2 weeks.

Folkhälsomyndigheten is in the process of removing *several thousand* positive test results from their data, so far mostly from July. https://twitter.com/Sven_Roman/status/1299766976957554690

The problem is, the vast majority of websites that collate data for different countries do so using the data reported by authorities on a given day. If, eg, Sweden had 1000 new cases reported today, but they also removed 1000 cases from July, this would give a net change of 0

You'll thus get wrong interpretations like @Sven_Roman has done here.

Because most of the sites look at rolling averages, usually 7 days, this means it will take a week *after the last correction is made* before these sites report Sweden's data correctly again.

Now, also to be clear - there's nothing nefarious going on here. Neither @Folkhalsomynd or the various reporting websites are doing anything wrong. A problem has been found and the data is being corrected. That's fine, it happens.

Just *always* be aware of such things when interpreting data, and if it seems odd, go looking for reasons why!

/end