Content
@
0 reply
0 recast
1 reaction
Geoff Golberg
@geoffgolberg
👀 Usernames have been added! 👀 https://warpcast.com/geoffgolberg/0xf501170e
7 replies
5 recasts
19 reactions
Jacob
@jrf
what parameters does this consider other than followers?
1 reply
1 recast
1 reaction
Geoff Golberg
@geoffgolberg
Specific to the influence rankings, eigenvector centrality algorithm effectively captures an account’s ability to propagate information within a network (i.e. it is focused solely on followers) That said, Social Forensics uses a bunch of other data attributes when investigating CIB (coordinated inauthentic behavior)
2 replies
1 recast
2 reactions
Jacob
@jrf
i think the warpcast algorithm is currently the driving force of network reach (including channels, push notifications, recasts, and replies), so i don't think we can read too much into the table.. though we do all love a good, clean table
1 reply
1 recast
0 reaction
Geoff Golberg
@geoffgolberg
following/followers, according to Dan, are THE most important thing re: WC home feed algo https://warpcast.com/dwr.eth/0xfe877a6b
1 reply
0 recast
0 reaction
Jacob
@jrf
... sure, you'll see casts from users you follow, that's a given, but in practice i think things tend to break out and be seen by more people when there's more engagement, i.e. replies, recasts, quotes
1 reply
0 recast
1 reaction
Geoff Golberg
@geoffgolberg
My point is that reducing eigenvector centrality algorithm-driven influence rankings to a “good, clean table” doesn’t align with the criteria rankings Dan has provided re: WC home feed algo
1 reply
0 recast
0 reaction
Jacob
@jrf
well, this is an extreme example, but if, for instance, you follow 100 people, and 90 don't cast one day, and the other 10 do, you'll see the content from those 10 and likely also the replies of the people who engage with the content maybe that goes without saying?
1 reply
0 recast
1 reaction
