## load packages
library("cluster")

# set working directory for module and check
setwd("C:/Users/ufrei/Desktop/WD-module")
getwd()

#load the MeanMeanEarPhenotypes dataset and look at the structure
data <- read.csv("MeanMeanEarPhenotypes.csv", header = T)
str(data)
#remove the first column (Pedigree) of the dataset by substting
datap <- subset(data, select=-(Pedigree:Type))
head(datap)

###### Generating the distance matrix
# calculate an Euclidean distance matrix and round the data to three digits
Dist <- dist(datap, method="euclidean")
round(Dist, digits=3)
Dist


#### Hierarchical Clustering using different methods: single, complete, average=UPGMA

x<- hclust(Dist, method="single")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering using single")

x<- hclust(Dist, method="complete")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering using complete")

x<- hclust(Dist, method="average")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering using average")


##### Playing around by excluding variables form theoriginal dataset
head(datap)

EarL <- subset(data, select=EarL)
Dist <- dist(EarL, method="euclidean")
round(Dist, digits=3)
x<- hclust(Dist, method="average")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering based on EarL")

ca300KWt <- subset(data, select=ca300KWt)
Dist <- dist(ca300KWt, method="euclidean")
round(Dist, digits=3)
x<- hclust(Dist, method="average")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering based on ca300KWt")

RowNo <- subset(data, select=RowNo.)
Dist <- dist(RowNo, method="euclidean")
round(Dist, digits=3)
x<- hclust(Dist, method="average")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering based on RowNo")

EarDT <- subset(data, select=EarDT)
Dist <- dist(EarDT, method="euclidean")
round(Dist, digits=3)
x<- hclust(Dist, method="average")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering based on EarDT")

Cobdim <-subset(data, select=EarL:CobDT)
Dist <- dist(Cobdim, method="euclidean")
round(Dist, digits=3)
x<- hclust(Dist, method="average")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering based on cob dimensions")

Cob <-subset(data, select=c("RowNo.", "EarL", "CobWt"))
Dist <- dist(Cob, method="euclidean")
round(Dist, digits=3)
x<- hclust(Dist, method="average")
plot(x, labels = data$Pedigree, hang = -0.1, main = "Hierarchical Clustering based on selected variables")