C19 update

I have made some progress on the plots. It turns out that running the plots with the current Johns Hopkins data, which is real time, introduces an issue. Specifically, the current day point is a partial value and pushes the data fit downward. Dropping the current day partial value is a better way to go and these plots use that approach. Even though the plot insert text says Apr09 today, it is based on data through Apr08. Subtle point, but this can have a significant effect on the model fit when it is run forward for many time steps.

On the cumulative plot there is now an insert text showing the estimated total.

The daily plot has only one y axis which is daily C19 deaths. The red dots are computed by differencing on the cumulative Johns Hopkins data. This process loses the first data point since the first difference is on the second day, that is why the time axis begins here at 1/23/2020. The model fit curve is computed as an analytical derivative on the fitted logistic function. Insert text has been added showing the day/value of the model fit peak.

Q&A
Q: Wow - don't know whether to hope you're wrong or right. Thanks for your work. MP
A: I think it will be pretty accurate till the world daily death count peaks. Based on the few countries past their peak, the downside is not as steep as the upside. May have to tweak a new model to handle that. In other words, my plot is about the best we can hope for....