Coronavirus Forecast of Active Cases (U.S.)
For the definition of "Active Case" see the Entry Page.
23 June 2020: U.S.
Orange curve: Original forecast I made based on a mere guess on March 21.
There is a considerable discrepancy for the past week or two between Johns Hopkins data and the date from https://www.worldometers.info/coronavirus/country/us/, a dip for several days before a new rise. I do not have an explanation for this difference.
UPDATE FOR SWITZERLAND: 23 June 2020
23 June 2020: WASHINGTON STATE
I generated the forecast model shown above for the U.S. with the orange bars on March 21, 2020. Reported data is shown with the blue bars. This is only for the U.S. I looked at other random predictions from statistical models on the web and based mine on a mere hunch, using two parameters to create the graph: (1) Maximum active cases of around 1/2 million (slightly more), (2) A fairly rapid decrease in active cases to a very low value by the end of June 2020. I have no statistical justification or special insight for these parameters. I generated the graph by entering round numbers into Excel and seeing how the graph took shape. The graph doesn't look very plausible overall to me, but the point was to put something together that I could compare with reported data on a daily basis.
I had to reconstruct (kludge) most of the early data from a variety of sites I listed on the entry page. From March 21 on I use only the Johns Hopkins data, which I enter daily usually between 6 and 7 pm, but always between 4 and 9 pm. I figure the daily fluctuations that occur by not recording the JH data at the same time every day are insignificant over a longer period of time. I manually enter data from the JH site into my spreadsheet, so it probably contains a few "scribal" errors now and then.
I will update periodically to see how the real numbers compare with my "forecast."