Each country has it's own date for when "testing came online", and it seems like starting from then might make things more comparable, despite differences in testing capacity and strategy. Another approach might be to on start looking at the data after the number of cases is > 100.