[Feature request] Provide raw data of app reports and JHU/RKI reports to allow time series analysis about dka

[corneliusroemer]

I'm stupid, I found the raw data in your repo. Here we go :)
https://github.com/micb25/dka/blob/master/data_CWA/correlation_CWA_JHU.csv

If I find time I do a time series lag analysis to find the proper denoised proportion, which would be one of the two key plots of your dashboard in my view. The other being how many are reporting, maybe even in one plot. One more thing that could be of big interest is the number of days submitted - since that is an indicator of proper usage (if only a few dates uploaded, it suggests most people don't have it on most of the time).

[corneliusroemer]

Ok, I ran a quick time lag cross correlation over the RKI, JHU and CWA series and it turns out that the optimal lag between RKI and CWA is 0, but between JHU and CWA it's 1 (so JHU should be taken from previous day with current CWA day), same goes for JHU and RKI by the way.
Regarding the meaning of report day for CWA, is the day the day on which the report was made by the individual, or the day on which the keys were broadcast to everyone (which should be one day after). I presume it's the latter, explaining the 1 day delay observed in cross correlation.