Njal
@cryptonjal
Are we casting better, or are spammy accounts casting less? In general, spammy accounts are casting less. - Number of casts from spammy accounts without a reply are shrinking - Interestingly enough, the number of casts from spammy accounts (0) with a reply is rising lately. Not sure if it's related to a rise in rising crypto prices, but could be. - Replies from not spammy (2) accounts seems to be more stable. - Also based on "quality engagement" in my Dune dashoard, spammy accounts seems to be more active from May 11. At that time, altoins were doing pretty well. Also, based on what I wrote this morning: it seems that the best way to get rid of spammy replies is a bear market. But, at the moment the number of active spammy accounts isn't declining. I removed casts from maybe spammy (1), and no labels to avoid time-outs in Dune. Here's the complete dashboard: https://dune.com/nhejyht/farcaster-dashboard
5 replies
3 recasts
12 reactions
Oliver
@timewarp
I link you and @mvr together. Sorry if that's not the case. Have either of you considered making a channel that requires all members to be a lvl 2 spam label? Making it more apparent through a badge or membership seems like the right type of incentive
2 replies
0 recast
3 reactions
Pichi
@pichi
I don’t think any of the mod tools are currently pulling Neynar score or spam label (they are highly correlated) but I could be wrong. @sahil would know for sure.
1 reply
0 recast
3 reactions
Njal
@cryptonjal
@mvr did great work with his mini apps, and Dune data. I like his work. I wouldn't recommend using one measurement, and use it strictly without purpose. Every measurement has false positives and negatives. You could combine the spam label with other measurements like the Neynar score, and/or your own judgement. If I remembered well, the Neynar score measures the network effect. Combining multiple measurement allows you to be more strict or less strict. After all, less popular channels doesn't need (many) controls to get rid of spammy comments at this moment.
0 reply
0 recast
1 reaction