How is the conclusion wrong? If there is a cohort of developers in Japan that do use SO, then the statistics are still valid even if all developers in Japan don't use it.
Because the traffic is based on % of all vim visits. This means, if less Japanese visit, the share they have in the total % of all vim visits is also less.