Local social is one of the most attempted startup ideas of the 2010s. It is quite challenging because network effects are hard to reach at the local scale, and you need significant moderation to prevent bullying.
Yes, I see scaling as the biggest problem. Instead of developing one service for potentially 7 billion users, you're building 10 services for maybe 1 to 10 million users each. It would probably come down to patching together open-source solutions like Keycloak, Immich, FreeTube, Flarum, etc. I find atproto, on which Bluesky/Eurosky are based, particularly interesting.
I'm not so sure whether bullying is really a big problem. In such a regional environment, it makes little sense to appear "anonymous." You might not want your name and address publicly available, but if you comment on the popular parties and clubs, neighborhood happenings, etc., people who know you in real life will naturally recognize you. Ultimately, it comes down to how the local population operates.
A big issue is the NextDoor effect. The most prolific posters are going to be people who saw a black person walking around and are startled by it and posting about it. Sane people see that and leave.
I'm not so sure whether bullying is really a big problem. In such a regional environment, it makes little sense to appear "anonymous." You might not want your name and address publicly available, but if you comment on the popular parties and clubs, neighborhood happenings, etc., people who know you in real life will naturally recognize you. Ultimately, it comes down to how the local population operates.