12

I am creating a heat map using ggplot(), and would like to utilize the 3 color scheme of scale_fill_gradient2(). I've found, however that the middle color is too broad and tends to display some of my data negatively (using "black" for example). Is it possible to define the range that is considered "mid," to make it more narrow? If not, is there a better way that I may do so?

Data Set:

structure(list(var1 = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 
5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 
5L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 
6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 
7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 
7L), .Label = c("A", "B", "C", "D", "E", "F", "G"), class = "factor"), 
var2 = structure(c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 
11L, 12L, 13L, 14L, 15L, 16L, 17L, 18L, 19L, 20L, 21L, 22L, 
23L, 24L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 
13L, 14L, 15L, 16L, 17L, 18L, 19L, 20L, 21L, 22L, 23L, 24L, 
1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 14L, 
15L, 16L, 17L, 18L, 19L, 20L, 21L, 22L, 23L, 24L, 1L, 2L, 
3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 
16L, 17L, 18L, 19L, 20L, 21L, 22L, 23L, 24L, 1L, 2L, 3L, 
4L, 5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 16L, 
17L, 18L, 19L, 20L, 21L, 22L, 23L, 24L, 1L, 2L, 3L, 4L, 5L, 
6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 16L, 17L, 18L, 
19L, 20L, 21L, 22L, 23L, 24L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 
8L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 16L, 17L, 18L, 19L, 
20L, 21L, 22L, 23L, 24L), .Label = c("a", "b", "c", "d", 
"e", "f", "g", "h", "i", "j", "k", "l", "m", "n", "o", "p", 
"q", "r", "s", "t", "u", "v", "w", "x"), class = "factor"), 
corr = c(0.039063517, -0.012531832, 0.096287532, -0.156156609, 
-0.097044878, 0.144426494, 0.102142979, 0.061426893, 0.051079225, 
0.271860908, 0.156812951, 0.259456277, -0.121838722, -0.157440078, 
-0.037827967, -0.01929319, -0.108895665, -0.066815122, 0.254285337, 
0.12688199, -0.064394035, 0.00112601, 0.173774179, 0.01179886, 
0.105171013, 0.088559148, 0.033584364, -0.368075609, -0.272671354, 
-0.456557935, -0.441008229, -0.118498286, -0.309056047, 0.051624421, 
0.087594347, -0.0264506, 0.081249807, -0.194887615, -0.135397719, 
-0.078688964, -0.059544125, -0.065410158, 0.211446055, 0.027338504, 
-0.06185598, -0.007720807, 0.092997248, -0.177812491, 0.133226267, 
0.075247459, 0.04586679, -0.37972917, -0.254410003, -0.447919321, 
-0.426264017, -0.150347417, -0.270786314, 0.143483685, 0.230384468, 
-0.012297462, 0.096957204, -0.134348613, -0.056239035, -0.038059581, 
-0.040273741, -0.131126698, 0.222754865, 0.067883188, -0.154724805, 
-0.076366467, 0.152747678, -0.160657826, 0.104652439, 0.029599007, 
-0.02194356, -0.349623751, -0.228046427, -0.363148102, -0.326649794, 
-0.160199177, -0.253621489, 0.242711785, 0.242808683, 0.005937022, 
0.120849407, -0.131358491, 0.004558011, -0.056873299, 0.048745675, 
-0.147360978, 0.204314713, 0.033110005, -0.154644375, -0.095999454, 
0.202067265, -0.182694112, 0.071256125, 0.062783632, 0.078250678, 
-0.348208521, -0.237899511, -0.235431913, -0.249280276, -0.129802124, 
-0.084153729, 0.201859004, 0.292353092, -0.000270168, 0.130668616, 
-0.101519227, -0.024695776, -0.024373331, -0.05976031, -0.07861482, 
0.150316096, 0.100494694, -0.107507982, -0.06543542, 0.220813403, 
-0.159374132, 0.139677124, 0.005974473, 0.011484774, -0.423918147, 
-0.286712875, -0.286979217, -0.2730134, -0.095014441, -0.134216671, 
0.215551068, 0.432571228, 0.029945066, 0.178996421, -0.114005412, 
-0.098205883, 0.009661061, -0.07163473, 0.010722592, 0.175801779, 
0.170582102, -0.199407024, -0.15933157, 0.246575881, -0.084929336, 
0.160642274, 0.025941888, -0.058676465, -0.297996589, -0.180504241, 
-0.236672516, -0.228203539, -0.181349045, -0.091893083, 0.277145017, 
0.38119709, 0.019554381, 0.18725277, -0.019416086, 0.000417258, 
0.134560252, 0.059330986, -0.091994882, 0.11807573, 0.175219716, 
-0.169272253, -0.094240107, 0.180532375, 0.04322002)), .Names = c("var1", "var2", "corr"), class = "data.frame", row.names = c(NA, -168L))

Code using reshape2, ggplot2, and scales:

sample.m <- melt(sample)
sample.m$var2 <- as.character(sample.m$var2)
sample.m$var2 <- factor(sample.m$var2, levels=unique(sample.m$var2),ordered=TRUE)
sample.m$var1 <- as.character(sample.m$var1)
sample.m$var1 <- factor(sample.m$var1, levels=unique(sample.m$var1),ordered=TRUE)
p <- ggplot(sample.m, aes(var2,var1)) + geom_tile(aes(fill=value), colour="transparent") +    scale_fill_gradient2(low="cyan", mid="black", high="red", midpoint=0, limits=c(-0.6,0.6))
base_size <- 9
p + theme_bw() + labs(x = "", y = "") + scale_x_discrete(expand = c(0,0)) + theme(legend.position = "bottom", axis.ticks = element_blank(), axis.text.x = element_text(size = base_size * 0.8, angle = 310, hjust = 0, colour = "black"), axis.text.y = element_text(size = base_size * 0.8,colour = "black"))
diaferiaj
  • 315
  • 1
  • 5
  • 13

3 Answers3

22

You can try scale_fill_gradientn and the values argument. From ?scale_fill_gradientn:

if colours should not be evenly positioned along the gradient this vector gives the position (between 0 and 1) for each colour in the colours vector. See rescale for a convience function to map an arbitrary range to between 0 and 1.

Thus, resolution of the colour scale for values close to zero may be increased by using suitable numbers in values = rescale(...).

scale_fill_gradientn(colours = c("cyan", "black", "red"),
                       values = scales::rescale(c(-0.5, -0.05, 0, 0.05, 0.5)))

enter image description here

Henrik
  • 65,555
  • 14
  • 143
  • 159
  • 1
    +1, didn't realize `colour` and `values` needn't be the same length, though in retrospect thinking about how this all works it makes sense. – BrodieG Feb 13 '14 at 15:43
  • Excellent! I adjusted to `(-0.5, -0.1, 0, 0.1, 0.5)` and it looks phenomenal. – diaferiaj Feb 13 '14 at 16:00
  • 1
    @Henrik If colour are values are not the same length, then, how does colours get distributed across values? in your solution does the black colour goes from -0.05 to 0.05? – chang02_23 Dec 15 '17 at 20:09
  • @chang02_23 Under the hood, [scale_fill_gradientn](https://github.com/tidyverse/ggplot2/blob/master/R/scale-gradient.r) calls `scales::gradient_n_pal`, which in turn call `scales::colour_ramp`. Please check the code and the help text for these functions (and for `scales::rescale`) and play around with easy toy examples, and I think you will get a better feeling for how they work. Cheers. – Henrik Dec 15 '17 at 21:38
10

Try replacing your scale_fill_gradient2 with:

scale_fill_gradientn(values=c(1, .6, .5, .4, 0), colours=c("red", "#770000", "black", "#007777", "cyan"))

You can adjust how narrow you want the black part to be by changing the .6 and .4 values (and also potentially the colors around "black"). This produces:

enter image description here

Compare to what you had:

enter image description here

BrodieG
  • 51,669
  • 9
  • 93
  • 146
1

As of 21.03.2023, you can simply set the midpoint to some value in scale_fill_gradient2

scale_fill_gradient2(midpoint = 1000)

You can also select the midpoint color

scale_fill_gradient2(midpoint = 1000, low = "green", mid = "blue", high = "red")
NelsonGon
  • 13,015
  • 7
  • 27
  • 57