This is a collection of R scripts for performing analysis on Are.na.
It uses the spider package to crawl Are.na and build a graph of channel connections.
From there you can build a text corpus of channel IDs with MCMC and then use that for topic modeling using word2vec.