dendro.r

data <- iris

d <- dist(scale(iris[, 1:4]))
h <- hclust(d, "ward")

data$ORDER <- order(h$order)
data$HEIGHT <- 0
data$LEVEL <- 0
data$POINTS <- 1

combine <- function(x, weights) {
  if (is.factor(x)) {
    if (length(unique(x)) == 1) x[1] else NA
  } else {
    weighted.mean(x, weights, na.rm = TRUE)
  }
}
combine_rows <- function(df) {
  combined <- as.data.frame(llply(df, combine, weights = df$POINTS))
  combined$POINTS <- sum(df$POINTS)
  combined
}

for (i in 1:nrow(h$merge)) {  
  newr <- combine_rows(data[as.character(-h$merge[i,]),])
  newr$HEIGHT <- h$height[i]
  newr$LEVEL <- i
  rownames(newr) <- as.character(-i)

  data <- rbind(data, newr)
}

data$node <- (as.numeric(rownames(data)) < 0) + 0

leaves <- subset(data, node == 0)
nodes <- subset(data, node == 1)
  
# < 0 = observation, > 0 = cluster
merge <- h$merge
merge[merge > 0] <- merge[merge > 0] + nrow(leaves)
merge[merge < 0] <- abs(merge[merge < 0])

lines <- data.frame(rbind(
  cbind(nodes[, ], data[merge[, 1], ]),
  cbind(nodes[, ], data[merge[, 2], ])
))

ggplot(, aes(ORDER, POINTS)) + 
  geom_point(data = leaves) + 
  geom_point(data = nodes, size = 0.5)  + 
  geom_segment(aes(xend=ORDER.1, yend=POINTS.1, group=1), data=lines) +
  xlab(NULL) + 
  scale_y_continuous("Number of points in cluster")