Farcaster

Piggybacking of @geoffgolberg and his influence metrics, I did a very light first pass on a statistical test whether or not the pagerank influence score correlates with Neynar score. I'm using a bootstrapping technique on the Spearman Rank for statistically minded people, which is below. Long story short, the distribution of these numbers is so close to 0 for casters with over 1000 followers.

With the data I have now, it appears there really isn't much of a link between how much influence one has from the Neymar engineered scores, and the graph itself. It could mean that maybe we can add pagerank scores to Neynar scores, but I'm not sure how that's actually calculated anyways (would love to learn). 

It might suggest mixing the pagerank in with Neynar scores to get some better signals. I could see a world where a pagerank, World Id, and Neynar score along with other metrics are all blended to fight the bots.

CC @dwr.eth and @rish

Working on Farcaster and Warpcast. Longer thoughts at https://dwr.email

building farcaster infra @ /neynar 🪐 | casting @ /rish

I like running, poker and oysters | Founder, Social Forensics (SocialForensics.com) | Berkeley & UPenn | Learning to listen

Also for giggles, on why I think blending the Neynar score with PageRank is free product improvement. 😅 

This *has* to be a bug in my code cause this is wild if true.

Slight statement as well:

This is just a first pass on the data with the social graph around early last year. I'll keep downloading more data until Neynar kicks me out I guess, or until I can't fit anymore data, but it seems at least promising to ask about the relationship between Neynar scores and the social graph signals at the moment.

nfa dyor