in R

# 
library(igraph)
data <- read.csv("http://commres.net/wiki/_media/r/socialnetworkdata.csv", header=T)
head(data)
str(data)

y <- data.frame(data$first, data$second)
head(y)

# net <- graph.data.frame(y, directed=T)
net <- graph_from_data_frame(y, directed=T)
head(net)
net

V(net) # vertex in net data
who.net <- V(net) # 52/52 vertices
data.frame(who.net)

E(net) # edge info in net data
rel.net <- E(net)
rel.net # output 290/290 edges 

# 52 by 290 data set

V(net)$degree <- degree(net)
degree.net <- data.frame(degree(net))
V(net)$degree


str(who.net)
who.net$name

hist(V(net)$degree)

set.seed(222)
plot(net,
     vertex.color = 'lightblue',
     vertext.size = 2,
     edge.arrow.size = 0.1,
     vertex.label.cex = 0.8)

plot(net,
     vertex.color = rainbow(52),
     vertex.size = V(net)$degree*0.3,
     edge.arrow.size = 0.1,
     layout=layout.fruchterman.reingold)
# layout layout.fruchterman.reingold
# layout.graphopt
plot(net,
     vertex.color = rainbow(52),
     vertex.size = V(net)$degree*0.8,
     edge.arrow.size = 0.1,
     layout=layout.graphopt)

plot(net,
     vertex.color = rainbow(52),
     vertex.size = V(net)$degree*0.4,
     edge.arrow.size = 0.1,
     layout=layout.kamada.kawai)

betweenness(net)
degree(net)
closeness(net)
hits_scores(net)$hub          # outlinks
hits_scores(net)$authority    # inlinks

hs <- hits_scores(net)$hub          # outlinks
as <- hits_scores(net)$authority    # inlinks



set.seed(123)
plot(net,
     vertex.size=hs*30,
     main = 'Hubs',
     vertex.color = rainbow(52),
     edge.arrow.size=0.1,
     layout = layout.kamada.kawai)

set.seed(123)
plot(net,
     vertex.size=as*30,
     main = 'Authorities',
     vertex.color = rainbow(52),
     edge.arrow.size=0.1,
     layout = layout.kamada.kawai)


net <- graph_from_data_frame(y, directed = F)
cnet <- cluster_edge_betweenness(net)

# cluster_fast_greedy(net)
# cluster_fluid_communities(net, 2)
# cluster_infomap(net)
# cluster_label_prop()
# cluster_spinglass()
# cluster.distribution()
# components(net, "strong")

infomap <- cluster_infomap(net)
# c.infomap <- cluster_infomap(net)

plot(cnet,
     net,
     vertex.size = 10,
     vertex.label.cex = 0.8)

plot(infomap,
     net,
     vertex.size = 10,
     vertex.label.cex = 0.8)

############################################################
# clustering in R
# clusterdata.csv
orgs <- read.csv("http://commres.net/wiki/_media/r/clusterdata.csv", header=T)
str(orgs)
head(orgs)

pairs(orgs[2:9])

plot(Fuel_Cost ~ Sales, data = orgs)
with(orgs, text(Fuel_Cost ~ Sales, 
                labels=orgs$Company,
                pos=4))

# normalization (standardization)
z <- orgs[,-c(1,1)] # remove the first column
means <- apply(z, 2, mean)
sds <- apply(z, 2, sd)
means
sds

nor <- scale(z, center=means, scale=sds)
nor 

distance <- dist(nor)
distance

orgs.hclust = hclust(distance)

plot(orgs.hclust)
plot(orgs.hclust,
     labels = orgs$Company,
     main='Default from hclust')
plot(orgs.hclust,
     hang=-1, 
     labels=orgs$Company, 
     main='Default from hclust')


orgs.hclust.average<-hclust(distance, method="average") 
plot(orgs.hclust.average, hang=-1,
     labels=orgs$Company, 
     main='hclust average')

plot(orgs.hclust.average, hang=-1,
     labels=orgs$Company, 
     main='hclust average')
rect.hclust(orgs.hclust.average, k=4, border='red')

member.by.3 <- cutree(orgs.hclust, 3)
table(member.by.3)
member.by.3

member.by.5 <- cutree(orgs.hclust, 5)
table(member.by.5)
member.by.5


aggregate(nor,list(member.by.3),mean)
aggregate(nor,list(member.by.3),sd)

wss <- (nrow(nor)-1)*sum(apply(nor,2,var))
for (i in 2:20) 
  wss[i] <- sum(kmeans(nor, centers=i)$withinss)
plot(1:20, wss, 
     type="b", 
     xlab="Number of Clusters", 
     ylab="Within groups sum of squares")


# kmeans clustering 
set.seed(123)
kc<-kmeans(nor, 3)
kc

# install.packages("cluster")
library(cluster)

ot<-nor
datadistshortset <- dist(ot, method = "euclidean")
hc1 <- hclust(datadistshortset, method = "complete" )
pamvshortset <- pam(datadistshortset, 4, diss = FALSE)
clusplot(pamvshortset, shade = FALSE, 
         labels=2, 
         col.clus="blue",
         col.p="red", 
         span=FALSE,
         main="Cluster Mapping",
         cex=1.2)

# kmeans
# install.packages("factoextra")
library(factoextra) 
k2 <- kmeans(nor, centers = 3, nstart = 25)
str(k2)

fviz_cluster(k2, data = nor)
fviz_nbclust(nor, kmeans, method = "wss")
fviz_nbclust(nor, kmeans, method = "silhouette")

gap_stat <- clusGap(nor, FUN = kmeans, 
                    nstart = 25, K.max = 10, B = 50)
fviz_gap_stat(gap_stat)