#### Packages install 
##install.packages("igraph")
##install.packages("igraphdata")
##install.packages("network") 
##install.packages("networkD3")
##install.packages("circlize")
install.packages("sand")


##Load packages
library("igraph")
library("igraphdata")
library("network") 
library("networkD3")
library("circlize")
library(sand)

## Set working directory
##setwd("C:/.....)


# Reproducible layout
set.seed(42)

## At the heart of graph visualization is the graph layout, i.e., the placement of vertices 
#and edges in space.

### Random dataset

# random graph Erdos-Renyi model with nodes = 20 and density = 0.2 --undirected links

g <- sample_gnp(50, 0.1, directed=FALSE)
?sample_gnp
edge_density(g) ##density of g
g

#  let sample graph from another model preferential attachment model

g.pa <- sample_pa(50, 1, directed=FALSE)
?sample_pa
edge_density(g.pa) ##density of g
g.pa

#let  use a well-known dataset aidsblog

data("aidsblog")
summary(aidsblog)

# adjacency matrices for graphs g e aidsblog

gm <- get.adjacency(g, sparse=F)

aidsm <- get.adjacency(aidsblog, sparse=F)
 

## different layouts for graph

igraph.options(vertex.size=5, vertex.label=NA, vertex.color="orange", edge.size=15, edge.color="grey", edge.arrow.size=0.2)

par(mfrow=c(1, 3))

# random layout
plot(g, layout=layout_randomly)
title("random graph")
plot(g.pa, layout=layout_randomly)
title("pref. attachment")
plot(aidsblog, layout=layout_randomly)
title("Blog Network")

#  circular layout
plot(g, layout=layout.circle)
title("random graph")
plot(g.pa, layout=layout.circle)
title("pref. attachment")
plot(aidsblog, layout=layout.circle)
title("Blog Network")

# Fruchterman & Reingold layout
plot(g, layout=layout_with_fr)
title("random graph")
plot(g.pa, layout=layout_with_fr, vertex.label = V(g)$name)
title("pref. attachment")
plot(aidsblog, layout=layout_with_fr)
title("Blog Network")

# MDS layout 
plot(g, layout=layout_with_mds)
title("random graph")
plot(g.pa, layout=layout_with_mds)
title("pref. attachment")
plot(aidsblog, layout=layout_with_mds)
title("Blog Network")

# Kamada-Kawai layout

plot(g, layout=layout_with_kk)
title("random graph")
plot(g.pa, layout=layout_with_kk)
title("pref. attachment")
plot(aidsblog, layout=layout_with_kk)
title("Blog Network")

dev.off()

distances(aidsblog, mode = "all")
## Compute node degree and use it to set node size in the graph

## igraph::degree to use the degree fuction in igraph package
par(mar=c(0.5,0.5,0.5,0.5)) ### set margins

deg_g <- igraph::degree(g, mode = "all", normalized = T)

V(g)$size <- deg_g*100
V(g)$name <- 1:vcount(g)
V(g)$label.color <- "blue"
plot(g, layout=layout_with_fr, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Random graph. Node size = node degree centrality", cex.main=1)

## Compute node closeness and use it to set node size

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 
clos_g <- igraph::closeness(g, mode = "all", normalized = T)
V(g)$size <- clos_g*30
V(g)$label.color <- "blue"
plot(g, layout=layout_with_fr, vertex.label = V(g)$name,  vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Random graph. Node size = node closeness centrality", cex.main=1)

## Compute node betweenness and use it to set node size

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 
?betweenness

betw_g <- igraph::betweenness(g, directed = F, normalized = T)
V(g)$size <- betw_g*100
V(g)$label.color <- "blue"
plot(g, layout=layout_with_fr, layout=layout_with_fr, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Random graph. Node size = node betweenness centrality", cex.main=1)

## Compute node eigenvector and use it to set node size

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 
eigen_g <- eigen_centrality(g)
V(g)$size <- eigen_g$vector*10
V(g)$label.color <- "blue"
plot(g, layout=layout_with_fr, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Random graph. Node size = node eigenvector centrality", cex.main=1)
?layout_with_fr

## preferential attachment - undirected graph

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 

plot(g.pa)
title(main="Graph layout default", cex.main=0.8)
plot(g.pa, layout=layout_randomly, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6)
title("Randomly place vertices _PA", cex.main=0.8)
plot(g.pa, layout=layout_with_fr, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6)
title(main="Fruchterman-Reingold layout -PA", cex.main=0.8)
plot(g.pa, layout=layout_with_kk, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6)
title(main="Kamada-Kawai layout -PA", cex.main=0.8)
plot(g.pa, layout=layout_in_circle, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6)
title(main="Graph layout circle -PA", cex.main=0.8)
plot(g.pa, layout=layout_with_mds, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6)
title(main="Graph layout multidimensional scaling -PA", cex.main=0.8)
dev.off()

## Compute node degree and use it to set node size
par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 
deg_g.pa <- igraph::degree(g.pa, mode = "all", normalized = T)
V(g.pa)$name <-  1:vcount(g.pa)
V(g.pa)$label.color <- "blue"
V(g.pa)$size <- deg_g.pa*100
plot(g.pa, layout=layout_with_fr, vertex.label = V(g.pa)$name, vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Pref. Attach. graph. Node size = node degree centrality", cex.main=1)


#### Compute node closeness and use it to set node size

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 
clos_g.pa <- closeness(g.pa, mode = "all", normalized = T)


V(g.pa)$size <- clos_g.pa*10

plot(g.pa, layout=layout_with_fr, vertex.label = V(g.pa)$name, vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Pref. Attach. graph. Node size = node closeness centrality", cex.main=1)

#### Compute node betweenness and use it to set node size

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 
betw_g.pa <- betweenness(g.pa, directed = F, normalized = T)
V(g.pa)$size <- betw_g.pa*10
plot(g.pa, layout=layout_with_fr, vertex.label = V(g)$name, vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Pref. Attach. graph. Node size = node betweenness centrality", cex.main=1)

#### Compute node eigenvector and use it to set node size

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 
eigen_g.pa <- eigen_centrality(g.pa)
V(g.pa)$size <- eigen_g.pa$vector*10
plot(g.pa, layout=layout_with_fr, vertex.label = V(g.pa)$name, vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Pref. Attach. graph. Node size = node eigenvector centrality", cex.main=1)


##  random graph - directed graph

AB <- aidsblog
par(mfrow=c(2,3))

plot(AB)
title(main="Graph layout default", cex.main=0.8)
plot(AB, layout=layout_randomly)
title("Randomly place vertices _AB", cex.main=0.8)
plot(AB, layout=layout_with_fr)
title(main="Fruchterman-Reingold layout -AB", cex.main=0.8)
plot(AB, layout=layout_with_kk)
title(main="Kamada-Kawai layout -AB", cex.main=0.8)
plot(AB, layout=layout_in_circle)
title(main="Graph layout circle -AB", cex.main=0.8)
plot(AB, layout=layout_with_mds)
title(main="Graph layout multidimensional scaling -AB", cex.main=0.8)
dev.off()

# if you want to freeze node position...

coord <-  layout_with_fr(AB) 

## Compute node indegree and use it to set node size
par(mar=c(0.5,0.5,0.5,0.5)) ### set margins 
indeg_AB<- igraph::degree(AB, mode = "in", normalized = T)
V(AB)$Indegsize <- indeg_AB*50
plot(AB, layout=coord, vertex.size=V(AB)$Indegsize) 
title(main="Aidsblog. Node size = node indegree centrality", cex.main=1)

# with  the outdegree

outdeg_AB<- igraph::degree(AB, mode = "out", normalized = T)
V(AB)$Outdegsize <- outdeg_AB*50
plot(AB, layout=coord, vertex.size=V(AB)$Outdegsize) 
title(main="Aidsblog. Node size = node outdegree centrality", cex.main=1)


#with the overall degree

deg_AB<- igraph::degree(AB, mode = "all", normalized = T)
V(AB)$degsize <- deg_AB*50
plot(AB, layout=coord, vertex.size=V(AB)$degsize) 
title(main="Aidsblog. Node size = node overall degree centrality", cex.main=1)


#  bipartite graphs often are laid out with the two sets of vertices running
# across opposing rows (or down opposing columns)

plot(g.bip, layout=-layout.bipartite(g.bip)[,2:1], vertex.size=30, vertex.shape=ifelse(V(g.bip)$type,"rectangle", "circle"), vertex.color=ifelse(V(g.bip)$type, "red", "cyan"))

#################################################################################
#decorating graphs

### Karate undirected graph

data(package="igraphdata")
data(karate)
class(karate) # graph object
table(V(karate)$Faction)

karatem <- get.adjacency(karate, sparse=F)
dim(karatem)
class(karatem) # matrix
edge_density (karate)

# Plot undecorated first

par(mar=c(0.5,0.5,0.5,0.5))### set margins
igraph.options(vertex.size=10)
l1 <- layout.kamada.kawai(karate)
plot(karate, layout=l1, vertex.label=V(karate))
title(main="undecorated plot karate", cex.main=0.8)

# Now decorate, starting with labels
V(karate)$label <- sub("Actor ", "", V(karate)$name)
# Two leaders get shapes different from club members.
V(karate)$shape <- "circle"
V(karate)[c("Mr Hi", "John A")]$shape <- "rectangle"
# Differentiate two factions by color.
V(karate)[Faction == 1]$color <- "red"
V(karate)[Faction == 2]$color <- "dodgerblue"
# Vertex area proportional to vertex strength
# (i.e., total weight of incident edges).
V(karate)$size <- 4*sqrt(graph.strength(karate))
V(karate)$size2 <- V(karate)$size * .5
# Weight edges by number of common activities among karate club members
E(karate)$width <- E(karate)$weight
# Color edges by within/between faction.
F1 <- V(karate)[Faction==1]
F2 <- V(karate)[Faction==2]
E(karate)[ F1 %--% F1 ]$color <- "pink"
E(karate)[ F2 %--% F2 ]$color <- "lightblue"
E(karate)[ F1 %--% F2 ]$color <- "yellow"
# Offset vertex labels for smaller points (default=0).
V(karate)$label.dist <-
  ifelse(V(karate)$size >= 10, 0, 0.75)
# Plot decorated graph, using same layout.
plot(karate, layout=l1)
title(main="decorated plot karate", cex.main=0.8)
V(karate)$Faction

#############################################################
## one-mode network - undirected graph  
## dataset florentine families (Padgett data) --parental relation

library(network)
data(package="network")
data("flo")
class(flo) # is a matrix


### trasform matrix in igraph object
flo.adiacency <- graph_from_adjacency_matrix(flo, mode="undirected")

class(flo.adiacency) # is a graph object
edge_density(flo.adiacency)

### node labels
V(flo.adiacency)$label <- sub("Actor ", "", V(flo.adiacency)$name)

### all graphs layouts for florentine families

par(mar=c(1.5,1.5,1.5,1.5)) ### set margins
igraph.options(vertex.size=10, vertex.label=V(flo.adiacency)$name, vertex.label.dist=0.8, vertex.color="orange", vertex.label.cex=0.8)
par(mfrow = c(2,3)) 
plot(flo.adiacency, layout=layout_nicely(flo.adiacency))
title(main="Florentine graph layout nicely", cex.main=0.8)
plot(flo.adiacency, layout=layout_randomly(flo.adiacency))
title(main="Florentine graph layout randomly place vertices", cex.main=0.8)
plot(flo.adiacency, layout=layout_with_fr(flo.adiacency))
title(main="Florentine graph layout Fruchterman-Reingold", cex.main=0.8)
plot(flo.adiacency, layout=layout_with_kk(flo.adiacency))
title(main="Florentine graph layout Kamada-Kawai", cex.main=0.8)
plot(flo.adiacency, layout=layout_in_circle(flo.adiacency))
title(main="Florentine graph layout circle", cex.main=0.8)
plot(flo.adiacency, layout=layout_with_mds(flo.adiacency))
title(main="Florentine graph layout multidimensional scaling", cex.main=0.8)
dev.off()


### to save coordinates and plot nodes on the same coordinates
coordfr<-layout_with_fr(flo.adiacency)
plot(flo.adiacency, layout=coordfr)

#### Compute node degree and use it to set node size

## florentine families - undirected graph

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins
deg_flo <- igraph::degree(flo.adiacency, mode = "all", normalized = T)
V(flo.adiacency)$size <- deg_flo*100
V(flo.adiacency)$label.color <- "blue"
plot(flo.adiacency, layout=layout_with_kk(flo.adiacency), vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="Florentine graph. Node size = node degree centrality", cex.main=1)

####statistics

# Degree distribution
par(mar=c(4,4,4,4)) ### set margins
dd <- degree.distribution(flo.adiacency, cumulative=T, mode="all")
plot(dd, pch=19, cex=0.8, col="orange", xlab="Degree", ylab="Cumulative")



# Heatmap of the network matrix - sociomatrix representation

palf <- colorRampPalette(c("white", "red")) 
heatmap(flom[, 16:1], Rowv = NA, Colv = NA, col = palf(2), 
        scale="none", margins=c(6,6) )
##title(main="heatmap plot florentine", cex.main=0.8)

# Heatmap of the network matrix - sociomatrix representation
#### reordering adiacency matrix according to clustering - geodesic distances 
geoflo<-distances(flo.adiacency, v = V(flo.adiacency), to = V(flo.adiacency), mode ="all", weights = NULL, 
                  algorithm = "automatic")
geoflo[geoflo==Inf]<-15
geoflo<-as.dist(geoflo)
hclust_geoflo<-hclust(geoflo, method = "complete", members = NULL)
heatmap(flom,Rowv=hclust_geoflo$order,Colv="Rowv", scale="none")
###title(main="heatmap plot florentine", cex.main=0.8)


### Directed and weighted graphs - one-mode network 

data(package="igraphdata")
data(UKfaculty)
class(UKfaculty) # graph object
table(V(UKfaculty)$Group)
edge_density (UKfaculty)

UKfacultym <- get.adjacency(UKfaculty, sparse=F)

## node color based on attribute data

par(mar=c(0.5,0.5,0.5,0.5)) ### set margins

igraph.options(vertex.size=6, vertex.label=NA, vertex.label.dist=1, vertex.label.cex=0.8, edge.color="lightblue", edge.arrow.size=0.4)
par(mfrow = c(2,3)) 
E(UKfaculty)$width <- E(UKfaculty)$weight/3
V(UKfaculty)$color <- V(UKfaculty)$Group
plot(UKfaculty, layout=layout_nicely(UKfaculty))
title(main="UKfaculty directed graph layout nicely", cex.main=0.8)
plot(UKfaculty, layout=layout_randomly(UKfaculty))
title("UKfaculty directed graph randomly place vertices", cex.main=0.8)
plot(UKfaculty, layout=layout_with_fr(UKfaculty))
title(main="UKfaculty directed graph Fruchterman-Reingold layout", cex.main=0.8)
plot(UKfaculty, layout=layout_with_kk(UKfaculty))
title(main="UKfaculty directed graph Kamada-Kawai layout", cex.main=0.8)
plot(UKfaculty, layout=layout_in_circle(UKfaculty))
title(main="UKfaculty graph layout circle", cex.main=0.8)
plot(UKfaculty, layout=layout_with_mds(UKfaculty))
title(main="UKfaculty directed graph layout multidimensional scaling", cex.main=0.8)

dev.off() 

### separate figures with different layouts
### different shapes and colors for groups

par(mar=c(0.5,0.5,0.5,0.5))  ### set margins
E(UKfaculty)$width <- E(UKfaculty)$weight/3
V(UKfaculty)$color <- V(UKfaculty)$Group
V(UKfaculty)[V(UKfaculty)$Group=="1"]$shape <- "rectangle"
V(UKfaculty)[V(UKfaculty)$Group=="2"]$shape <- "circle"
V(UKfaculty)[V(UKfaculty)$Group=="3"]$shape <- "square"
V(UKfaculty)[V(UKfaculty)$Group=="4"]$shape <- "sphere"
plot(UKfaculty, layout=layout_nicely(UKfaculty), vertex.size=5, vertex.label=NA, edge.color="lightblue", edge.arrow.size=0.6)
title(main="UKfaculty directed graph layout nicely", cex.main=0.8)

par(mar=c(0.5,0.5,0.5,0.5))  ### set margins
E(UKfaculty)$width <- E(UKfaculty)$weight/3
V(UKfaculty)$color <- V(UKfaculty)$Group
V(UKfaculty)[V(UKfaculty)$Group=="1"]$shape <- "rectangle"
V(UKfaculty)[V(UKfaculty)$Group=="2"]$shape <- "circle"
V(UKfaculty)[V(UKfaculty)$Group=="3"]$shape <- "square"
V(UKfaculty)[V(UKfaculty)$Group=="4"]$shape <- "sphere"
plot(UKfaculty, layout=layout_randomly(UKfaculty), vertex.size=5, vertex.label=NA, edge.size=15, edge.color="lightblue", edge.arrow.size=0.4)
title("UKfaculty directed graph randomly place vertices", cex.main=0.8)

par(mar=c(0.5,0.5,0.5,0.5))  ### set margins
E(UKfaculty)$width <- E(UKfaculty)$weight/3
V(UKfaculty)$color <- V(UKfaculty)$Group
V(UKfaculty)[V(UKfaculty)$Group=="1"]$shape <- "rectangle"
V(UKfaculty)[V(UKfaculty)$Group=="2"]$shape <- "circle"
V(UKfaculty)[V(UKfaculty)$Group=="3"]$shape <- "square"
V(UKfaculty)[V(UKfaculty)$Group=="4"]$shape <- "sphere"
plot(UKfaculty, layout=layout_with_fr(UKfaculty), vertex.size=5, vertex.label=NA, edge.size=15, edge.color="lightblue", edge.arrow.size=0.4)
title(main="UKfaculty directed graph Fruchterman-Reingold layout", cex.main=0.8)

par(mar=c(0.5,0.5,0.5,0.5))  ### set margins
E(UKfaculty)$width <- E(UKfaculty)$weight/3
V(UKfaculty)$color <- V(UKfaculty)$Group
V(UKfaculty)[V(UKfaculty)$Group=="1"]$shape <- "rectangle"
V(UKfaculty)[V(UKfaculty)$Group=="2"]$shape <- "circle"
V(UKfaculty)[V(UKfaculty)$Group=="3"]$shape <- "square"
V(UKfaculty)[V(UKfaculty)$Group=="4"]$shape <- "sphere"
plot(UKfaculty, layout=layout_with_kk(UKfaculty), vertex.size=5, vertex.label=NA, edge.size=15, edge.color="lightblue", edge.arrow.size=0.4)
title(main="UKfaculty directed graph Kamada-Kawai layout", cex.main=0.8)

par(mar=c(0.5,0.5,0.5,0.5))  ### set margins
E(UKfaculty)$width <- E(UKfaculty)$weight/3
V(UKfaculty)$color <- V(UKfaculty)$Group
V(UKfaculty)[V(UKfaculty)$Group=="1"]$shape <- "rectangle"
V(UKfaculty)[V(UKfaculty)$Group=="2"]$shape <- "circle"
V(UKfaculty)[V(UKfaculty)$Group=="3"]$shape <- "square"
V(UKfaculty)[V(UKfaculty)$Group=="4"]$shape <- "sphere"
plot(UKfaculty, layout=layout_in_circle(UKfaculty), vertex.size=5, vertex.label=NA, edge.size=15, edge.color="lightblue", edge.arrow.size=0.4)
title(main="UKfaculty graph layout circle", cex.main=0.8)

par(mar=c(0.5,0.5,0.5,0.5))  ### set margins
E(UKfaculty)$width <- E(UKfaculty)$weight/3
V(UKfaculty)$color <- V(UKfaculty)$Group
V(UKfaculty)[V(UKfaculty)$Group=="1"]$shape <- "rectangle"
V(UKfaculty)[V(UKfaculty)$Group=="2"]$shape <- "circle"
V(UKfaculty)[V(UKfaculty)$Group=="3"]$shape <- "square"
V(UKfaculty)[V(UKfaculty)$Group=="4"]$shape <- "sphere"
plot(UKfaculty, layout=layout_with_mds(UKfaculty), vertex.size=5, vertex.label=NA, edge.size=15, edge.color="lightblue", edge.arrow.size=0.4)
title(main="UKfaculty directed graph layout multidimensional scaling", cex.main=0.8)


#### Compute node indegree and use it to set node size

## UKfaculty friendship relationship  directed and weighted graph

par(mar=c(0.5,0.5,0.5,0.5)) ## set margins
indeg_uk <- igraph::degree(UKfaculty, mode = "in", normalized = T)
##outdeg_uk <- degree(UKfaculty, mode = "out", normalized = T)
V(UKfaculty)$size <- indeg_uk*50
##V(UKfaculty)$size <- out_uk*100
V(UKfaculty)$color <- V(UKfaculty)$Group
E(UKfaculty)$width <- E(UKfaculty)$weight/3
plot(UKfaculty, layout=layout_with_fr(UKfaculty), vertex.label.dist=0, vertex.label.cex=0.6, edge.arrow.size=0.4) 
title(main="UKfaculty graph. Node size = node indegree centrality", cex.main=1)

####statistics

# Degree distribution
par(mar=c(4,4,4,4)) ## set margins
dd <- degree.distribution(UKfaculty, cumulative=T, mode="all")
plot(dd, pch=19, cex=0.8, col="orange", xlab="Degree", ylab="Cumulative Frequency")

#### Compute node closeness and use it to set node size

par(mar=c(0.5,0.5,0.5,0.5)) ## set margins
clos_uk <- closeness(flo.adiacency, mode = "all", normalized = T)
V(UKfaculty)$size <- clos_uk*10
V(UKfaculty)$color <- V(UKfaculty)$Group
E(UKfaculty)$width <- E(UKfaculty)$weight/3
plot(UKfaculty, layout=layout_with_fr(UKfaculty), vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="UKfaculty graph. Node size = node closeness centrality", cex.main=1)

#### Compute node betweenness and use it to set node size
par(mar=c(0.5,0.5,0.5,0.5)) ## set margins
betw_uk <- betweenness(UKfaculty, directed = T, normalized = T)
V(UKfaculty)$size <- betw_uk*100
V(UKfaculty)$color <- V(UKfaculty)$Group
E(UKfaculty)$width <- E(UKfaculty)$weight/3
plot(UKfaculty, layout=layout_with_kk(UKfaculty), vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="UKfaculty graph. Node size = node betweenness centrality", cex.main=1)

#### Compute node eigenvector and use it to set node size
par(mar=c(0.5,0.5,0.5,0.5)) ## set margins
eigen_uk<- eigen_centrality(UKfaculty)
V(UKfaculty)$size <- eigen_uk$vector*50
V(UKfaculty)$color <- V(UKfaculty)$Group
E(UKfaculty)$width <- E(UKfaculty)$weight/3
plot(UKfaculty, layout=layout_with_kk(UKfaculty), vertex.label.dist=0, vertex.label.cex=0.6) 
title(main="UKfaculty graph. Node size = node eigenvector centrality", cex.main=1)


####heatmap UKfaculty

colnames(UKfacultym) <- V(UKfaculty)$name
rownames(UKfacultym) <- V(UKfaculty)$name
palf <- colorRampPalette(c("red", "white")) 
heatmap(UKfacultym[,81:1], Rowv = V(UKfaculty)$Group, Colv = "Rowv", col = palf(20), 
        scale="none", margins=c(6,6))
##title(sub="heatmap UKfaculty", cex.sub=1)

###heatmap by Group 
heatmap(UKfacultym[order(V(UKfaculty)$Group), ], col = palf(10), labRow= NA, labCol= NA)


##################################################################
### ego-network visualization

#the so-called egocentric network visualizations commonly used in the
#social network literature, which show the vertex, its immediate neighbors, and all edges among them.

k.nbhds <- graph.neighborhood(karate, order=1)

# the neighborhoods pertaining to the head instructor (Mr Hi, vertex 1) and administrator (John A, vertex 34) are the largest.

sapply(k.nbhds, vcount)

k.1 <- k.nbhds[[1]]
k.34 <- k.nbhds[[34]]

par(mfrow=c(1,2))
plot(k.1, vertex.label=NA, vertex.color=c("red", rep("lightblue", 16)))
plot(k.34, vertex.label=NA, vertex.color=c(rep("lightblue", 17), "red"))
dev.off()

##################################################################
####two-mode network

# Random bipartite graph
inc <- matrix(sample(0:1, 50, replace = TRUE, prob=c(2,1)), 10, 5)
gtwo <- graph_from_incidence_matrix(inc)
plot(gtwo, layout = layout_as_bipartite, vertex.color=c("blue","red")[V(gtwo)$type+1])


### two-mode networks Davis Data

davis<-read.table("Davis.txt",header = T, row.names = 1)
davis<-as.matrix(davis)
dim (davis)
davis.affiliation <- graph.incidence(davis)
is_bipartite(davis.affiliation)

par(mar=c(1.5,1.5,1.5,1.5)) ##set margins
igraph.options(vertex.size=8, vertex.label.dist=1, vertex.label.cex=0.3)
plot(davis.affiliation, layout = layout_as_bipartite, vertex.label=V(davis.affiliation)$name,
     vertex.color=c("orange","blue")[V(davis.affiliation)$type+1])
title(main="Simple two-row layout for bipartite graphs", cex.main=0.8)



### Circos graphs
par(mar=c(0.5,0.5,0.5,0.5)) ##set margins
chordDiagram(davis)



### from two-mode network to two one-mode networks 
## how to derive two one-mode networks from a two mode network

bipart<-bipartite.projection(davis.affiliation)

## one.mode network nodes x nodes
donneDavis<-bipart$proj1
class(donneDavis)
donneDavis_m <- get.adjacency(donneDavis, sparse=F)
class(donneDavis_m)
dim(donneDavis_m)

## ## one.mode network events x events
eventiDavis<-bipart$proj2
class(eventiDavis)
eventiDavis_m <- get.adjacency(eventiDavis, sparse=F)
class(eventiDavis_m)
dim(eventiDavis_m)

## Interactive plotting with tkplot 

## R and igraph offer interactive plotting capabilities
## (mostly helpful for small networks)

tkid <- tkplot(flo.adiacency) #tkid is the id of the tkplot
l <- tkplot.getcoords(tkid) # grab the coordinates from tkplot
plot(flo.adiacency, layout=l)

plot(flo.adiacency, layout=layout_with_fr)
tkplot(flo.adiacency, layout=layout_with_fr)

##################################################################
### network from package networkd3

## florentine families example

flo_data<-as_data_frame(flo.adiacency, what="edges")
flo_data=flo_data[,1:2]
simpleNetwork(flo_data, width = 400, height = 250)
?simpleNetwork
## UKfaculty example1

UK_data<-as_data_frame(UKfaculty)
simpleNetwork(UK_data, Source = "from", Target = "to", nodeColour = "orange", width = 400, height = 250)