Louvain

Louvain is a desirable algorithm in this case almost purely due to the lack of parameters.
We can see some downsides when looking closely at the largest clustered communities, however. For example, in 2015, the gang-bang node is loosely attached to seemingly separate communities such as club and real-amateur, where we might expect separation.

2015

make_plot(year_filter = 2015, type = "louvain", min_edges = 10, min_group_size = 7, width_shrink_factor = 7)

2016

make_plot(year_filter = 2016, type = "louvain", min_edges = 10, min_group_size = 5, width_shrink_factor = 7)

2017

make_plot(year_filter = 2017, type = "louvain", min_edges = 10, min_group_size = 5, width_shrink_factor = 7)

2018

make_plot(year_filter = 2018, type = "louvain", min_edges = 10, min_group_size = 7, width_shrink_factor = 7)

2019

make_plot(year_filter = 2019, type = "louvain", min_edges = 15, min_group_size = 7, width_shrink_factor = 150)

2020

make_plot(year_filter = 2020, type = "louvain", min_edges = 70, min_group_size = 7, width_shrink_factor = 200)