![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
undyingkingThis is quite a neat idea -- these guys have applied a PageRank-like algorithm to LJ to determine a user's LiveRank, using friend and friend-of relationships as the analogy of links.
There's a discussion here of the need to exclude broadcast-type accounts, etc.
NB this is not live data -- it's based on a dataset from late 2004. Doing it with live data would be prohibitively demanding.
This made me think of other such things. Clearly just counting friends and friends-of is a bit crude -- you'd want to assess comment activity and mentions of your LJ userid as well. These get very demanding very quickly, although not if you restrict the community to eg. one's own flist. Should allow you to eg. rank them in order of how friendly they really are! If I ever get around to doing it, I shall call it FriendRank...
There's a discussion here of the need to exclude broadcast-type accounts, etc.
NB this is not live data -- it's based on a dataset from late 2004. Doing it with live data would be prohibitively demanding.
This made me think of other such things. Clearly just counting friends and friends-of is a bit crude -- you'd want to assess comment activity and mentions of your LJ userid as well. These get very demanding very quickly, although not if you restrict the community to eg. one's own flist. Should allow you to eg. rank them in order of how friendly they really are! If I ever get around to doing it, I shall call it FriendRank...
