Also, for China & South Korea, you know they've got (relatively) good testing. For many other countries, it's best to treat these data as a lower bound. The real numbers could be much higher, meaning that fitting a straight line and deriving a growth rate might be inaccurate.