Looking at the pattern: if the statistic would end in autumn instead of summer, then iOS could also have come up on top.
Is this bad data? Is the market expanding & contracting so greatly that existing users are a minority of future marketshare?
You pay for results?
I have a lot of insight into the apps side of things, and Android is somewhat smaller than iOS in Japan, both on revenue and downloads.
Do you have a source for that?