heatmap in R with cell area proportional to an observation

Question

I want to produce heatmap in R where the cell area and color are governed by two columns. For example, in the data below, I want the area of every cell to be proportional to the "count" column and the color to be controlled by the "measure" column and put the "rank" label on each cell. How can I do this in R?

   Event    measure rank    Count
   A         20     1      2000
   B         15     2      1870
   C         10     3      1540
   D         5      4      1200

Answer 1

The type of plot you're looking for is called a treemap. I've edited the answer to include a table below the plot based on your comment.

library("treemap")
library("grid")
library("gridExtra")

mydat <- structure(list(Event = structure(1:4, .Label = c("A", "B", "C",
  "D"), class = "factor"), measure = c(20L, 15L, 10L, 5L), rank = 1:4,
  Count = c(2000L, 1870L, 1540L, 1200L)), .Names = c("Event",
    "measure", "rank", "Count"), class = "data.frame", row.names = c(NA, -4L))

# Setup the parent layout viewport
  layout_vp <- viewport(layout = grid.layout(nrow = 2, ncol = 1,
  widths = unit(c(1, 1), "npc"),
  heights = unit(c(0.6, 0.4), "npc")), name = "lay_vp")

# Specify the location of the top and bottom viewports for the tree plot and table
  tree_vp <- viewport(layout.pos.col = NULL, layout.pos.row = 1, name = "vp_top")
  table_vp <- viewport(layout.pos.col = NULL, layout.pos.row = 2, name = "vp_bottom",
  height = unit(0.1, "npc"))

# Create grob for table
  table_grob <- tableGrob(mydat, rows = NULL, vp = table_vp)

# Combining the treemap and table
  grid.newpage()
  pushViewport(layout_vp)
  pushViewport(tree_vp)
  treemap(mydat, index = "rank", vSize = "Count", vColor = "measure", type = "value",
      position.legend = "none", vp = tree_vp)
  popViewport()
  pushViewport(table_vp)
  grid.draw(table_grob)
  popViewport()

Giving