The Johns Hopkins University, Baltimore, publishes current counts of confirmed and recovered cases as well as deaths in the Corona outbreak for many countries. I wrote a little program which downloads the time-series values and displays them graphically in a TChart. I know that such programs can be found in the web at many places now, but having the data immediately available makes the program very agile. And there is an option to use a logarithmic scale for cumulative data which allows to detect when the currently exponential increase of cases begins to level out hopefully. And holding the CTRL key and left mouse button down, and dragging the mouse across data points allows to fit an exponential curve to the data points; the time for doubling the case counts as well as an estimage of the case count in 1 and 2 weeks are displayed in the status bar.
The attached screenshot compares the growth of confirmed cases for China/Hubei, South Korea, Iran, Italy and Germany. The line fitted to the exponential plot of Germany, extrapolated two weeks into the future, tells that there will be 400.000 confirmed cases then if the current measures fail (and if testing capacity will be sufficient to handle this).
The sources can be found on my github at
https://github.com/wp-xyz/corona.git and can be compiled with Laz trunk or 2.0.6 and FPC trunk or 3.0.4 (others combinations not tested). I am developing on Windows, but I checked the program to run on Linux (Mint) and Mac (Mojave), too. For access to the Johns Hopkins server, the OpenSSL library must be available on the system; for Windows users, the libeay32.dll and ssleay32.dll must be available in the exe folder with the correct bitness.