Graph theory is such a hack here. The people that you like (rate highly) should also be weighted more heavily when rating others. Consequently, you'll end up with an enjoyability score based on your social likes for any event with a Nostr attendee.
Discussion
That's interesting! How would that work in practice as UI/UX for example? Do we have literal ratings on profiles visible only to you? Because a follow may not necessarily mean you like that person.
Every like you attribute towards a pubkey is an endorsement of your total endorsements. Then the people that they like is a fractional endorsement of their likes.
I see. Makes sense. The only issue might be with people who use likes for acknowledgement instead of endorsement (like myself). I will often like to show that I saw your comment without the need to reply, but my like may not indicate that I like what you said. I know... confusing.
I think the point remains. Even those people should have much more reactions (likes and zaps) towards the people they actually like. You could, in theory, take this a little further with Sentiment Analysis but it might be too intrusive for the ethos of NOSTR.
As far as UX, it could be a small tag that I choose (only I see it). Different icon for three or so kinds of “liking” someone. The color of it could be a spectrum that parallels the “score”. Then the clients have those tags (for the users who choose to add them) and that allows to give back a prediction in the same form of icon and color back to user for the people they don’t know.
Ya I would much rather assign tags to people - someone I consider very like-minded, or someone who is a friend but may not be on the same wavelength with me, or others who are interesting but I may not have much in common with. Each should have their own weighted influence on my social sphere.
