# Download the data from url "www.url.csv"
data.df <- read.csv(url("www.url.csv"))

# Convert the data to matrix selecting specific rows
matrix.data <- as.matrix(data.df [1:3, ])
#rownames
row.names(matrix.data) <- data.df$new
# Create new vector
new_vector<- as.numeric(data.df$newvector == "x")
#PCA analysis

#creating clustering model
cluster_model <- dist(data)
#cutree() to cut the hierarchial tree using h or k values
#Linkage methods
complete_clust<- hclust(dist(data), method = "complete")
# scale data
data.scaled <- scale(data)
scaled.hierarchial.model<- hclust(dist(data.scaled), method = "complete")

# K means clustering with 
kmeans(dataset, centres = centre value, nstart = number of times)
#plot with plot function
# Initialize total within sum of squares error: wss
wss <- 0
# Look over 1 to 16 possible clusters
for (i in 1:16) {
  # Fit the model: fit.model
  fit.model <- kmeans(data, centers = i, nstart = 20, iter.max = 50)
  # Save the within cluster sum of squares
  wss[i] <- fit.model$tot.withinss
}

# plot function
plot(1:16, wss, type = "b", 
     xlab = "label x", 
     ylab = "label y")

# Select number of clusters
k <- 

# Build model with k clusters: k.model
k.model <- kmeans(data, centers = k, nstart = 20, iter.max = 50)


# Plot of variable1 vs. variable2 by cluster membership
plot(data[, c("variable1", "variable2")],
     col = k.model$cluster,
     main = paste("heading", k, "clusters"),
     xlab = "variable1", ylab = "variable2")