753PS-proposed-tutchanges

This is an in-progress list of proposed changes to the tutorials for 753PS.

• (Copyediting)
• Code cleanup to conform to tidyverse/google style guide, remove deprecated calls/functions
• Fix bugs like incorrect use of "attr" in frequently-used get.brokerage and elsewhere
• Tutorial 1
  • Add discussion of weights implementations in statnet (named) vs igraph (hard-coded)
  • Add discussion of directly accessing statnet x_stat$gal components to mirror igraph is_x functions
• Tutorial 2
  • Get clarity on igraph and edge weights in path length - not "At this point in time"?
  • Add discussion of statnet::isolates as an indicator of igraph is_connected
  • Add discussion of unconnected networks and issues presented (esp. wrt upcoming centrality measures), demonstrate with simple constructed example
  • Add code for detecting and self-connecting isolates in both statnet and igraph
• Tutorial 3
  • Mostly code cleanup, content OK
• Tutorial 4
  • Explain differences in eigenvalue values between statnet and igraph; provide code to get them in sync if possible. Get to the bottom of differences in the airports networks.
• Tutorial 5
  • Get to the bottom of the vast differences in closeness centrality between the libraries
  • Test and explain Gould-Fernandez
  • Substantiate " it is useful to look at the correlations between scores to understand what they are doing in relation to one another" - why? how?
• Tutorial 6
  • As in all previous tuts, continue to update/build improved nodeinfo dataframe
  • Correct and explain "hamming, correlation, hamming, or gamma" methods in equiv.fun="sedist"
  • Dendrograms: explain "Node 9 is clearly different compared to the other nodes."
  • Explain why some nodes are not bottom-aligned
  • Explain, don't just show "You may want to experiment with using different methods for the option equiv.clust(..., cluster.method=""), replacing the default cluster.method="complete" with an alternative such as single, average, or ward.D. How does changing the clustering method change which nodes are closest to each other?"
  • Explain vacant clusters in dendrograms
  • Investigate/explain why cluster partitions using fixed h don't match blockmodel partitions with the same h (flomarriage, week 6 hw)
  • "In practice, there are different approaches to identifying the best cutpoint to use as the basis for partitioning, and we explore those in more detail below." - don't see it