do some clustering and create feeds that are lenses into all the different subcommunities
-
-
-
this one's my fav suggestion so far
End of conversation
New conversation -
-
-
determine clout metric for each user, capture all high-clout individuals
-
find out who else is dumb enough to follow akira despite the extreme oscillations between sad boi mopery and faux-Landian edgy acc-lordism and form a support group for people in a codependent relationship with bad content
- End of conversation
New conversation -
-
-
chop wood, carry water
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Find and collect people who are secretly ingroup
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
bifurcate it further by attacking critical nodes
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Read a bunch of graph theory papers and never actually use it, most likely
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
1. Find the communities. 2. Find the trendmakers in each community. 3. Produce personally directed propaganda to change their alignments. (Roughly at random.)
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
cluster analysis
-
there are fancier "community detection" algos, too, that the bioinfo people developed. nodes are members of a mixture of communities instead of just one. could also be lazy and just do lda on the adjacency matrix
End of conversation
New conversation -
-
-
Ask people on Twitter what to do with it.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Look for triangles.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
look at some structural things
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Click on a cat video. Who cares?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
1/ Run a community-finding algorithm to determine tribes. 2/ Count the number of verified accounts in each tribe.
-
3/ Verified accounts may tend to follow each other even if they belong to different tribes, so it would be interesting to think of a way to avoid mislabeling
-
4/ Train a neural net to predict tribal affiliation based on profile photo, then construct a "typical" tribal profile
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.