User talk:Ndevln

Hi, could you please add sources to your edits in Q83889294? The numbers of infected people are very different between different sources (WHO, RKI, Johns Hopkins University) so this is important information.

Hi,

RKI and WHO are the same official and validated numbers, but the publishing date of the WHO numbers is lagging one day behind. I always use the RKI numbers, but sadly you have to use the internet wayback machine to get the previous numbers published by the RKI. I don't know why John Hopkins should be a valid source.

I can try to do it, but the German page is the only one with complete case numbers so far. The Wikidata UI is horrible, but maybe you can help me. Is there a way to create templates, so I can create a new entry with predefined fields and maybe also some predefined data?

The RKI has situation reports in PDF format with the URL https://www.rki.de/DE/Content/InfAZ/N/Neuartiges_Coronavirus/Situationsberichte/2020-03-13-de.pdf?__blob=publicationFile You just need to change the date in the URL to get other editions.

I'm quite new to Wikidata myself, so I fear I cannot help you about the template issue.

Somebody came and added the references. Additionally, everything is now moved back a day. Which is technically correct, but I didn't bother using the reporting date and not the actual case date. (Since every old entry had to be changed). It took 1 hour, which is insane for such small changes.

Also, the WHO and RKI had a change in how they report the data. With the new data we have a big artificial spike between 19/20.03. Now the WHO uses a different reporting deadline, which results in minor differences in the numbers. But overall it doesn't matter.

My primary concern was that 1) The data source has to be documented and 2) It has to be made sure that everything comes from the data source so that the numbers are comparable. Therefore, I'm happy about Velvet's changes, even if there are small problems with the WHO data.

For the numbers there is only one source and these are the local Public Health Offices. The different publisher only have different reporting deadlines, but they all come from the same truth. Since the sampling rate (daily) is much more frequent than the time from Infection to Symptoms (3-6 days) plus the time to get tested (another 3 days), it doesn't matter which source you use, and they can be mixed. I started documenting the data to run my own analyses and I thought I could as well do it publicly, so everyone can use it. And for me the spike is a minor nuisance, but in the end it doesn't matter for the data to be useful.

So I think all numbers are comparable and over 3 days they all tell the same story, even if you mix them.