Not really meant for self hosting, but have you looked at listenbrainz? The software and data is open and it’s not a commercial service.

I use organic maps and it works really well

By mass or by volume?

Storing it as a sparse graph should reduce the storage requirements drastically, since most edges wouldn’t exist.

They also don’t get access to all the data that I think is most invasive (federated or not). I expect my posts, comments, votes, and follows on a public forum to be public. I don’t expect which posts I open, which comments I read, and how long I view each one for (a much larger and more invasive pool of data) to be public, and that’s what I don’t want Meta to get. By not using threads, they don’t get that. By using threads (or any Meta product) they do get that, and they probably use it to shovel more ads in your face.

While I am a little cautious of the possibility of EEE, I feel like the majority of fediverse users are anti-corporation and relatively technically informed, and would anticipate any attempts to extinguish it would be poorly received and ineffective. (Edit: although I do think this argument is reasonable and haven’t really decided whether I think federating with threads is a good idea)

Either way, federating with threads won’t give them any non-public information, which is substantially better than if you used their products directly. The other information is there for anyone to grab, so it’s kind of weird to complain about them reading it. If you put up a sign in your yard, you wouldn’t complain about people who walk by reading it.