-
Here’s some simple code for pulling updated COVID-19 daily death data (or cases) from Johns Hopkins CSSE into @Stata and formatting as a time series: gist.github.com/lukestein/1aafbdf26cfa06e850ce6c167cd417cf This is national, except for region/province breakdowns for seven countries (not the US).
-
You’re now all set up to run to not-epidemiologically-sound regressions like: nbreg D.deaths c.date##i.location ib1.dow if (D.deaths > 10), irr Now just ignore std errs, speculate what’s going on Thursdays/Fridays, and you’re halfway to viral medium post fame!
-
Update: Apparently older versions of @Stata don’t do as well importing csv files with unacceptable variable names. (The problem here is that the Hopkins data calls longitude “long,” a reserved word in Stata). Updated with a suggested fix: