This is the kind of stuff that should be possible and valuable to study but requires TikTok, Facebook, etc to run the study since the data is all on their serve

This is the kind of stuff that should be possible and valuable to study but requires TikTok, Facebook, etc to run the study since the data is all on their servers. I expect that the analysis has been done internally and just not released since it’s bad (ie, social media use is higher among suicides). Maybe a whistleblower will leak this stuff. I think an analysis is possible by building off their typical drop off analysis. I think that because I expect all drop offs are likely analyzed to restart and avoid future drop offs. So I expect they have a way to classify people who died vs left the platform. In the US we release a “death master file” that has names, dob, ssn a few years after death. If Japan has something similar it can be run against account holder data to see who actually died. I don’t know privacy law in Japan, but in the US cause of death is confidential. So we just know that a kid died and would need to predict cause of death from social media activity. With this low number, they could be investigated manually to see how many children had accounts and what patterns exist in common.