Constructing a hexagonal heat-map with custom colors in each cell

Question

I would like to generate a hexagonal lattice heat-map in which each cell represents a group. Likewise, each cell would be a hexagon with a unique color (fill, set by a column color in the data-frame) value, and a saturation (alpha) value corresponding to continuous decimal values from a chemical concentration dateset.

I would like to use a standardized data format which would allow me to quickly construct figures based on standardized datasets containing 25 groups.

For example, a datasheet would look like this:

      structure(list(group = 1:25, color = c("red", "brown1", "hotpink1", 
      "orange", "indianred1", "magenta", "darkgoldenrod1", "goldenrod1", 
      "gold", "deeppink", "yellow", "darkseagreen1", "aquamarine", 
      "plum", "mediumorchid4", "olivedrab1", "limegreen", "thistle1", 
      "violetred", "green4", "mediumseagreen", "darkviolet", "lightseagreen", 
      "dodgerblue2", "deepskyblue4"), alpha = c(NA, NA, NA, NA, NA, 
      NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 
      NA, NA, NA, NA), x = c(1, 1.5, 1.5, 2, 2, 2, 2.5, 2.5, 2.5, 2.5, 
      3, 3, 3, 3, 3, 3.5, 3.5, 3.5, 3.5, 4, 4, 4, 4.5, 4.5, 5), y = c(3, 
      3.5, 2.5, 4, 3, 2, 4.5, 3.5, 2.5, 1.5, 5, 4, 3, 2, 1, 4.5, 3.5, 
      2.5, 1.5, 4, 3, 2, 3.5, 2.5, 3)), class = "data.frame", row.names = c(NA, 
      -25L))

A plot of this kind in which alpha = 1 for all groups might look like this:

Whereas plots of dataset1 and dataset2 (included below) would look like these, respectively:

I would like to use something simple, like hexbin(), but I haven't figured out how to get that to work for this application.

Dataset1:

  structure(list(group = 1:25, color = c("red", "brown1", "hotpink1", 
  "orange", "indianred1", "magenta", "darkgoldenrod1", "goldenrod1", 
  "gold", "deeppink", "yellow", "darkseagreen1", "aquamarine", 
  "plum", "mediumorchid4", "olivedrab1", "limegreen", "thistle1", 
  "violetred", "green4", "mediumseagreen", "darkviolet", "lightseagreen", 
  "dodgerblue2", "deepskyblue4"), alpha = c(1, 1, 0.5, 0.5, 0.2, 
  0.2, 0, 0, 0.3, 0.1, 1, 0, 0, 0, 0.7, 0, 0, 0, 0, 0, 0, 0, 0, 
  0.5, 0.9), x = c(1, 1.5, 1.5, 2, 2, 2, 2.5, 2.5, 2.5, 2.5, 3, 
  3, 3, 3, 3, 3.5, 3.5, 3.5, 3.5, 4, 4, 4, 4.5, 4.5, 5), y = c(3, 
  3.5, 2.5, 4, 3, 2, 4.5, 3.5, 2.5, 1.5, 5, 4, 3, 2, 1, 4.5, 3.5, 
  2.5, 1.5, 4, 3, 2, 3.5, 2.5, 3)), class = "data.frame", row.names = c(NA, 
  -25L))

Dataset2:

structure(list(group = 1:25, color = c("red", "brown1", "hotpink1", 
"orange", "indianred1", "magenta", "darkgoldenrod1", "goldenrod1", 
"gold", "deeppink", "yellow", "darkseagreen1", "aquamarine", 
"plum", "mediumorchid4", "olivedrab1", "limegreen", "thistle1", 
"violetred", "green4", "mediumseagreen", "darkviolet", "lightseagreen", 
"dodgerblue2", "deepskyblue4"), alpha = c(0.3, 0.5, 0.6, 0, 0.7, 
0, 0, 0, 0, 0, 0, 0.5, 0.3, 0, 0, 0, 0, 0.6, 0.8, 0.5, 0.7, 0.5, 
0.5, 0.7, 0.5), x = c(1, 1.5, 1.5, 2, 2, 2, 2.5, 2.5, 2.5, 2.5, 
3, 3, 3, 3, 3, 3.5, 3.5, 3.5, 3.5, 4, 4, 4, 4.5, 4.5, 5), y = c(3, 
3.5, 2.5, 4, 3, 2, 4.5, 3.5, 2.5, 1.5, 5, 4, 3, 2, 1, 4.5, 3.5, 
2.5, 1.5, 4, 3, 2, 3.5, 2.5, 3)), class = "data.frame", row.names = c(NA, 
-25L))

Out of curiosity, what's the application of this? It isn't a chart type I've come across. Also if you've tried anything already, it would be helpful to see it in the post as a starting point — camille, Dec 21 '21 at 17:16
@camille This is for expressing floral volatile profiles. It is my own design, I haven't been totally satisfied with other ways of expressing this kind of data. I have tried `hexbin(dat$x, dat$y)` to try to set up the frame, but no success there. Then I tried to just diy it in baseplot, (`plot(x = dat$x, y = dat$y, xlim = c(.5, 5.5), ylim = c(.5, 5.5, pch = )`) but there is no hexagonal `pch` shape — JKO, Dec 21 '21 at 17:37
These are different from your specific question, but might be helpful or interesting: https://stackoverflow.com/q/40479111/5325862, https://stackoverflow.com/q/24006361/5325862, https://www.siddix.us/2021/03/07/hexbin/ — camille, Dec 22 '21 at 16:24

score 2 · Accepted Answer · answered Dec 22 '21 at 01:52

If you're open to creating the plot in Python, the following approach would work:

import matplotlib.pyplot as plt
from matplotlib.patches import RegularPolygon
import numpy as np

data = {'group': np.arange(1, 26),
        'color': ["red", "brown", "hotpink", "orange", "indianred", "magenta", "darkgoldenrod", "goldenrod", "gold", "deeppink", "yellow", "darkseagreen", "aquamarine", "plum", "mediumorchid", "olivedrab", "limegreen", "thistle", "violet", "green", "mediumseagreen", "darkviolet", "lightseagreen", "dodgerblue", "deepskyblue"],
        'alpha': np.ones(25)}

fig, ax = plt.subplots()
ax.set_aspect('equal')
ax.axis('off')

ind = 0
N = 5
for x in np.arange(1, 2*N):
    num_y = N - abs(x - N)
    for y in range(N + num_y, N - num_y, -2):
        hexagon = RegularPolygon((x, y/np.sqrt(3)), numVertices=6, radius=2 / 3, orientation=np.pi/2,
                                 alpha=data['alpha'][ind],
                                 facecolor=data['color'][ind], edgecolor='k')
        ax.add_patch(hexagon)
        ax.text(x, y/np.sqrt(3), f"Group{data['group'][ind]}", color='black', ha='center', va='center')
        ind += 1
plt.autoscale(enable=True)
plt.show()

Ben · Answer 2 · 2021-12-29T16:28:04.550

This is not exactly the designed use of hexbin, but it may give you something to work with. You will need to modify your coordinates to have the desired placement of hexagons (see dataset below, based on Dataset2). In addition, the hexagons are rotated.

library(hexbin)
library(ggplot2)

ggplot(df, aes(x = x, y = y, label = paste("Group", group))) + 
  geom_hex(stat = "identity", 
           color = "black", 
           fill = df$color, 
           alpha = df$alpha, 
           show.legend = F) + 
  geom_text() +
  scale_x_continuous(limits = c(0,6)) + 
  scale_y_continuous(limits = c(0,9)) +
  theme_void()

Plot

Or alternatively, using stat_bin_hex and specifying the binwidth you could try the following. The fill and alpha are sorted by value of y here.

ggplot(df, aes(x = x, y = y, label = paste("Group", group))) + 
  stat_bin_hex(color = "black", 
               fill = df[order(df$y), "color"], 
               alpha = df[order(df$y), "alpha"], 
               binwidth = c(1, 1), 
               show.legend = F) + 
  geom_text() +
  scale_x_continuous(limits = c(0,6)) + 
  scale_y_continuous(limits = c(0,9)) +
  theme_void()

Plot

Data

df <- structure(list(group = 1:25, x = c(1, 1.5, 1.5, 2, 2, 2, 2.5, 
2.5, 2.5, 2.5, 3, 3, 3, 3, 3, 3.5, 3.5, 3.5, 3.5, 4, 4, 4, 4.5, 
4.5, 5), y = c(4.5, 5.4, 3.6, 6.25, 4.5, 2.75, 7.15, 5.4, 3.6, 
1.85, 8.05, 6.25, 4.5, 2.75, 1, 7.15, 5.4, 3.6, 1.85, 6.25, 4.5, 
2.75, 5.4, 3.6, 4.5), color = c("red", "brown1", "hotpink1", 
"orange", "indianred1", "magenta", "darkgoldenrod1", "goldenrod1", 
"gold", "deeppink", "yellow", "darkseagreen1", "aquamarine", 
"plum", "blue2", "olivedrab1", "limegreen", "thistle1", "violetred", 
"green4", "mediumseagreen", "darkviolet", "lightseagreen", "dodgerblue2", 
"deepskyblue4"), alpha = c(0.3, 0.5, 0.6, 0, 0.7, 0, 0, 0, 0, 
0, 0, 0.5, 0.3, 0, 0, 0, 0, 0.6, 0.8, 0.5, 0.7, 0.5, 0.5, 0.7, 
0.5)), row.names = c(NA, -25L), class = "data.frame")

@JKO It's a great question - I did look around and couldn't find a way to rotate the hexagons so they would be oriented as in the example. You can rotate the whole figure using `coord_flip()`. — Ben, Dec 29 '21 at 16:37

Constructing a hexagonal heat-map with custom colors in each cell

2 Answers2

Linked