What data? What data could even support the idea of a threshold of acceptable deviation from the national average distribution of political thought?
Honestly, I think that while there is a problem, you have conceptualized it entirely incorrectly. The problem is not with the diversity of distribution of political thought, the problem is national political duopoly and the electoral systems that support it, among the many manifestations of which (and far from the most important) is an absence of partisan competition in jurisdictions whose political center isn't very close to the national political center.
> If you think it is possible to have an inter-subjective conversation about whether SF is too extreme
It's not (and I think obviously not), unless you specific for what. If you specify that, it's may be that such a discussion is possible (and even, though less likely, possible that it is meaningful and not a distraction from more important considerations for that purpose.) But I’m not optimistic.