The data
source("./Import Scripts/Game of Thrones Interactions.R")
The dataset is the Game of Thrones Interactions dataset, with characters as nodes, and instances of characters being mentioned in the same vicinity are the edges.
Here is the best network plot I have managed to create with this data so far:
V(network_igraph)$size<-5
plot(network_igraph, layout_with_lgl(network_igraph), vertex.label=NA)
Degree Centrality
head(arrange(data.frame(degree(network_igraph)), desc(degree.network_igraph.)), 10)
degree.network_igraph.
Robert Baratheon 233
Eddard Stark 220
Tywin Lannister 213
Robb Stark 212
Jaime Lannister 209
Stannis Baratheon 209
the Others 196
Arya Stark 187
Tyrion Lannister 184
Cersei Lannister 181#top 10 nodes with most degrees
This is the table I made last time, of the people that are in the Top 10 in degree centrality. I will be looking at the top ten lists for each of the centralities we are exploring today and comparing them.
Closeness
close_data<-data.frame(closeness(network_igraph))
head(arrange(close_data, desc(close_data)),10)
closeness.network_igraph.
Euron Greyjoy 0.001718213
Catelyn Tully 0.001706485
Renly Baratheon 0.001703578
Aerys II Targaryen 0.001694915
Petyr Baelish 0.001694915
Tywin Lannister 0.001686341
Tommen Baratheon 0.001683502
Roose Bolton 0.001680672
Nymeria 0.001663894
Aegon I Targaryen 0.001661130This measure is vastly different from our original degree centrality list. There are very few characters that appear in both of these lists. As closeness is a measure of how far away the rest of the nodes are from the character, this difference could probably be attributed to these characters having more connections to the further out nodes than the ones in the degree centrality list.
Betweenness
between_data<-data.frame(betweenness(network_igraph))
head(arrange(between_data, desc(between_data)),10)
betweenness.network_igraph.
Tywin Lannister 1203.9623
Eddard Stark 950.6527
Robert Baratheon 908.6096
Catelyn Tully 895.6626
the Others 889.9581
Robb Stark 850.6766
Roose Bolton 847.2783
Arya Stark 744.7401
Joffrey Baratheon 731.9701
Renly Baratheon 730.9969This list looks much more similar to the degree centrality list. While the order is not quite the same- Robert Baratheon is dethroned and relegated to third place in this table- many of the names are familiar. As this calculation relies on the number of close connections, that degree centrality will effect this measure makes sense.
Eigenvector Centrality
eigen<-centr_eigen(network_igraph,directed=F)
data<-data.frame(V(network_igraph)$name, eigen$vector)
head(arrange(data, desc(data$eigen.vector)), 10)
V.network_igraph..name eigen.vector
1 Robert Baratheon 1.0000000
2 Eddard Stark 0.9730299
3 Jaime Lannister 0.9583349
4 Stannis Baratheon 0.9402949
5 Tywin Lannister 0.9387183
6 Robb Stark 0.9227364
7 Tyrion Lannister 0.8942695
8 Cersei Lannister 0.8807297
9 the Others 0.8738112
10 Catelyn Tully 0.8676705Once again, Robert Baratheon is king! He is the node that has the most influence over the other characters- as he should! The rest of the list is almost identical to the degree centrality list- with some slight shuffling around, of course- with the only true difference being Catelyn Tully’s presence in lieu of Arya Stark’s. Since eigenvector accounts for how well-connected the connections of a node are, this makes sense- Arya may do a lot of traveling and meeting people, but she does so mostly in the periphery of the story. Meanwhile, Catelyn is striving for power and connections- while she may not have met as many people as Arya, she tries to be in the thoughts of important individuals, which would boost her up onto this list.