Need to create a pipeline that can run and save results for hundreds of communities
[ ] import data for each community
[ ] use AIC to determine number of topics
[ ] run LDA for that topic number
[ ] save LDA results (overall topic compositions, topic relative proportions for each sample)
[ ] feed results into changepoint model
[ ] automate application of changepoint model and saving results for many communities
[ ] automate determining for each community: # of changepoints, periods of consistent community structure, periods of gradual change, abrupt transitions
Need to create a pipeline that can run and save results for hundreds of communities
(Moves over @skmorgane's issue from https://github.com/weecology/macro-ts-struct/issues/3)