Show HN: Twitter: PageRank algo and, KMeans to cluster, discover 'who_to_follow'(nbviewer.ipython.org)
nbviewer.ipython.org
Show HN: Twitter: PageRank algo and, KMeans to cluster, discover 'who_to_follow'
http://nbviewer.ipython.org/github/balaca/twitter_graph/blob/master/twitter_graph.ipynb
2 comments
interesting outputs at end, but what could be the ways to improve clustering performance. e.g. in the clustering process output, VC firms and Data Science communities are evident while other clusters aren't as evident.
yes, I'm trying to understand how meaningfulness of Clusters in the graph improve as I change KMeans parameters. I'd appreciate more views from the community