Jay [MOVED HERE] on Nostr: I think right now the biggest hurdle is wrapping my head around the Nostr network and ...
I think right now the biggest hurdle is wrapping my head around the Nostr network and how each kind is formatted and used. I'll eventually run into more weird things the more outbox relays I connect to, for sure.
My plan right now is to focus on aggregating text notes threads, reacts, etc. and calculating trust scores based on engagement.
- Expand from my npub's notes and collect every related event and store them in lmdb and relational metadata in neo4j.
- As events come in, I have to store new users and their relays and increment an engagement score between us depending on the event kind and frequency.
- Once that's done, I can compare the incoming and outgoing engagement scores between me and all my peers to find those that are most equal.
My hypothesis is that those should be the most recognizable to me and my most trusted contacts. I can re-run the process using those as seeds and calculate the second sphere of trust. Eventually, this can become a personal relay of sorts that pulls in my network. If you let it keep going, you could theoretically map the entire network.
My plan right now is to focus on aggregating text notes threads, reacts, etc. and calculating trust scores based on engagement.
- Expand from my npub's notes and collect every related event and store them in lmdb and relational metadata in neo4j.
- As events come in, I have to store new users and their relays and increment an engagement score between us depending on the event kind and frequency.
- Once that's done, I can compare the incoming and outgoing engagement scores between me and all my peers to find those that are most equal.
My hypothesis is that those should be the most recognizable to me and my most trusted contacts. I can re-run the process using those as seeds and calculate the second sphere of trust. Eventually, this can become a personal relay of sorts that pulls in my network. If you let it keep going, you could theoretically map the entire network.