Todo: sunday meeting

[jwzimmer-zz]

Julia

• Vis for each of the four centrality measures for comparison
• get list to phil of tropes by indices
• clustering?

Anyone

• Get all the subindices as well as tropes for these main 4

Phil

• clustering on tropes (per some index subset)

[nguyenhphilip]

added the subindexes dict with links to their tropes for the main 4 indexes: 9f1c7e3

[jwzimmer-zz]

Using the subindex dict from 9f1c7e3: 8bfa7b9

[jwzimmer-zz]

Trying to show when tropes link to tropes adds too many edges: 5913cf0

[jwzimmer-zz]

Looking at just the NarrativeTropes with subindices: 11e3826

[jwzimmer-zz]

Finally a readable size (doing each of the big four separately, and only showing top 100 tropes):
NarrativeTropes
110b2d5

[jwzimmer-zz]

Topical tropes with subindices, but only if the subindices link to top 100 most central tropes in order to keep the network legible: 9f5b647

[jwzimmer-zz]

Media tropes with subindices + top 1000 tropes: 6283a27

[jwzimmer-zz]

Genre tropes + subindices: fb09f94 and 497361d

[jwzimmer-zz]

^ In the above comments, made graphs and visualizations for the big four + their subindices from phil's dict in #25 (comment)

[nguyenhphilip]

Two visualizations for communities made using girvan-newman (remove most 'valuable' edge to find communities). Edges only link to tropes in the list used for community detection.

first viz using 75 nodes, community size = 5:

second viz using 100 nodes, community size = 6: