It’s impossible really to say. This was their official code citation:
Over the past few months, we have carefully recomputed historical votes on posts and comments to remove outdated, unnecessary rules.
I mean on the face of it, maybe they were telling the truth?
But they are a for profit corporation, and that year forward was when the enshittification really began. I guess I just have little reason to believe that they didn’t just alter the algorithm to make it look like there was more engagement than there was.
Is everyone 5 days behind LW? I don’t quite really grasp how exactly there is a 5 day lag, shouldn’t Lemmy be close to real time? I don’t fully understand what’s going on in those charts but it looks like the delay will be gone in a few weeks/months?
As far as hosting it on a different instance, this is just kinda a playful experiment. If it goes well anyone can copy the list and put it a version of it on their server.
I clicked 5 of the links in the lemmydirectory github/wiki list and all of them the last post was 1 year ago. That’s what kinda sucks about all the indexes is that they’re either the extremely popular communities which show up in the communities tab, or extremely dead communities. The goal is to get a list of extremely active smaller communities and without changing Lemmy source code myself, this is the next best thing I could think of to make a list like that.