问题描述
样本数据-
Country segments variable value
US A Kerosene 16.09
US B Kerosene 2.81
US C Kerosene -7.96
US D Kerosene 25.76
US E Kerosene 75.47
US A Petrol 26.82
US B Petrol 8.72
US C Petrol 2.07
US D Petrol 16
US E Petrol -11
US A Edible Oil 4.02
US B Edible Oil 10.6
US C Edible Oil 49.05
US D Edible Oil 16.07
US E Edible Oil 51.87
US A LPG -1.07
US B LPG 49.19
US C LPG 37.9
US D LPG 33.05
US E LPG 102
这是美国的数据集，就像我在文件中拥有其他14个国家的数据集一样.现在，我用分层方法制作了15个单独的热图.x轴是变量，y轴是细分市场和热图，每个国家/地区都有关注值.
This is the data set for US, like that I have dataset of 14 other countries in the file. Now a days I make 15 separate heatmap in hierarchical approach. The x-axis is the variable and y-axis the segments and the heatmap with concern values based on each country.
我当时的想法是与所有国家一起制作一张幻灯片.因此，最终结果是-y轴与段相同，x轴与国家/地区相同，每个像元是一个热图，在单个像元中具有四个不同的变量和相应的值.
What I was thinking is to make a single slide with all the countries as well. So the final result with be - y-axis same with the segments, x-axis with countries and each cell is a heat map with four different variable and corresponding value in single cell.
任何帮助都将是非常有用的-任何有关如何编码和制作具有四个不同值和颜色代码的堆图的单个单元格的想法.如下所示-这将是一个包含8个国家和4个细分市场以及4个细分市场的示例变量.
Any help will be of great- Any idea of how I can code and make a single cell to a heapmap with four different values and color code.Something like below - That will be an example of 8 countries and 4 segments with 4 variables.
推荐答案
由于您没有提供完整的可复制示例(或您已经尝试过的任何代码)，因此我制作了一些虚拟数据来说明您可能会怎么做它.
Since you didn't provide a full reproducible example (or any code you've already tried), I made up some dummy data to show how you might do it.
library(dplyr)
library(ggplot2)
fake.df <-
expand.grid(Country = c("US", "Canada", "Mexico"),
segments = LETTERS[1:5],
variable = c("Kerosene", "Petrol",
"Edible Oil", "LPG")) %>%
mutate(value = runif(length(Country), -20, 50))
fake.df.adj <-
fake.df %>%
mutate(xadj = ifelse(variable %in% c("Kerosene", "Edible Oil"), -0.33/2, 0.33/2),
yadj = ifelse(variable %in% c("Kerosene", "Petrol"), 0.33/2, -0.33/2),
xpos = as.numeric(factor(Country)) + xadj,
ypos = as.numeric(factor(segments)) + yadj)
mutant 代码在要绘制的每个中心点周围创建一个调整框.对于煤油"和食用油"，这些图块向左移动每个 Country 之间的空间的1/3(因此，每个图块的中心为-0.33/2)；其他的则向右移动.与y调整相同，但上下左右而不是左右.
The mutate code creates a box of adjustments around each center point you want to plot. For "Kerosene" and "Edible Oil", those tiles are shifted to the left by 1/3 of the space between each Country (so the center of each tile is -0.33/2); the others are shifted right. Same with the y adjustment, but up and down instead of left and right.
然后将因子( Country 和 segments )转换为数字位置，并将它们加在一起.
Then you convert your factors (Country and segments) to numeric positions and add them together.
fake.df.adj
Country segments variable value xadj yadj xpos ypos
1 US A Kerosene -11.8318607 -0.165 0.165 0.835 1.165
2 Canada A Kerosene 39.1181835 -0.165 0.165 1.835 1.165
3 Mexico A Kerosene 25.9354644 -0.165 0.165 2.835 1.165
4 US B Kerosene -19.3503525 -0.165 0.165 0.835 2.165
5 Canada B Kerosene -12.4459506 -0.165 0.165 1.835 2.165
6 Mexico B Kerosene 9.9017016 -0.165 0.165 2.835 2.165
7 US C Kerosene 37.9767562 -0.165 0.165 0.835 3.165
8 Canada C Kerosene 3.4687111 -0.165 0.165 1.835 3.165
9 Mexico C Kerosene 0.6280239 -0.165 0.165 2.835 3.165
10 US D Kerosene 43.4739464 -0.165 0.165 0.835 4.165
11 Canada D Kerosene 8.4671510 -0.165 0.165 1.835 4.165
12 Mexico D Kerosene 19.7357609 -0.165 0.165 2.835 4.165
13 US E Kerosene -2.6120700 -0.165 0.165 0.835 5.165
14 Canada E Kerosene -10.6013792 -0.165 0.165 1.835 5.165
15 Mexico E Kerosene -6.2059698 -0.165 0.165 2.835 5.165
16 US A Petrol 17.1079974 0.165 0.165 1.165 1.165
17 Canada A Petrol 46.8944785 0.165 0.165 2.165 1.165
18 Mexico A Petrol 9.3977815 0.165 0.165 3.165 1.165
19 US B Petrol 38.7547489 0.165 0.165 1.165 2.165
20 Canada B Petrol -14.4210703 0.165 0.165 2.165 2.165
21 Mexico B Petrol 32.5160861 0.165 0.165 3.165 2.165
22 US C Petrol -1.3750645 0.165 0.165 1.165 3.165
23 Canada C Petrol 17.0540527 0.165 0.165 2.165 3.165
24 Mexico C Petrol -16.8449931 0.165 0.165 3.165 3.165
25 US D Petrol 33.8465349 0.165 0.165 1.165 4.165
26 Canada D Petrol 43.7369153 0.165 0.165 2.165 4.165
27 Mexico D Petrol 32.2145640 0.165 0.165 3.165 4.165
28 US E Petrol -13.1811532 0.165 0.165 1.165 5.165
29 Canada E Petrol 46.1913082 0.165 0.165 2.165 5.165
30 Mexico E Petrol -7.5030316 0.165 0.165 3.165 5.165
31 US A Edible Oil 36.1643957 -0.165 -0.165 0.835 0.835
32 Canada A Edible Oil 26.9887728 -0.165 -0.165 1.835 0.835
33 Mexico A Edible Oil 9.2825143 -0.165 -0.165 2.835 0.835
34 US B Edible Oil 32.7455557 -0.165 -0.165 0.835 1.835
35 Canada B Edible Oil 9.0447777 -0.165 -0.165 1.835 1.835
36 Mexico B Edible Oil 21.5822081 -0.165 -0.165 2.835 1.835
37 US C Edible Oil 2.3906442 -0.165 -0.165 0.835 2.835
38 Canada C Edible Oil 25.2536855 -0.165 -0.165 1.835 2.835
39 Mexico C Edible Oil 5.0618632 -0.165 -0.165 2.835 2.835
40 US D Edible Oil 41.0292115 -0.165 -0.165 0.835 3.835
41 Canada D Edible Oil 8.4515698 -0.165 -0.165 1.835 3.835
42 Mexico D Edible Oil 11.9080938 -0.165 -0.165 2.835 3.835
43 US E Edible Oil 42.2518838 -0.165 -0.165 0.835 4.835
44 Canada E Edible Oil 25.6458033 -0.165 -0.165 1.835 4.835
45 Mexico E Edible Oil -19.3037443 -0.165 -0.165 2.835 4.835
46 US A LPG 25.8891215 0.165 -0.165 1.165 0.835
47 Canada A LPG -3.1028641 0.165 -0.165 2.165 0.835
48 Mexico A LPG 47.5884154 0.165 -0.165 3.165 0.835
49 US B LPG 11.3000701 0.165 -0.165 1.165 1.835
50 Canada B LPG 26.8041755 0.165 -0.165 2.165 1.835
51 Mexico B LPG -13.0209453 0.165 -0.165 3.165 1.835
52 US C LPG -5.5292231 0.165 -0.165 1.165 2.835
53 Canada C LPG 46.3101034 0.165 -0.165 2.165 2.835
54 Mexico C LPG 19.7622448 0.165 -0.165 3.165 2.835
55 US D LPG 8.5258791 0.165 -0.165 1.165 3.835
56 Canada D LPG 20.5856857 0.165 -0.165 2.165 3.835
57 Mexico D LPG -10.7589733 0.165 -0.165 3.165 3.835
58 US E LPG 14.8577291 0.165 -0.165 1.165 4.835
59 Canada E LPG -6.9590007 0.165 -0.165 2.165 4.835
60 Mexico E LPG -3.1651872 0.165 -0.165 3.165 4.835
fake.df.adj %>%
ggplot(aes(xpos, ypos, fill = value)) +
geom_tile(height = 0.33, width = 0.33) +
geom_text(aes(label = paste0(round(value, 1), "%"),
color = value < 30), show.legend = F) +
geom_text(data = fake.df.adj %>% filter(Country == "Mexico", segments == "E"),
aes(label = variable,
y = ypos + 1.75*yadj),
fontface = "bold") +
scale_fill_gradient2(labels = function(breaks) paste0(breaks, "%")) +
scale_color_manual(values = c("FALSE" = "white", "TRUE" = "black")) +
scale_x_continuous(breaks = 1:3, name = NULL,
labels = levels(fake.df$Country)) +
scale_y_continuous(breaks = 1:5, name = "Segments",
labels = levels(fake.df$segments)) +
theme_classic()
现在，您可以使用 geom_tile 进行绘制，将图块的高度和宽度设置为0.33.
Now you can plot using geom_tile, setting the height and width of the tiles to be 0.33.
我在每个图块的顶部添加了标签，四舍五入到一个小数点后一位，并带有一个％符号.为了使它们在高值的黑暗背景填充下能很好地显示，我将标签的颜色美感作为逻辑表达式，然后进一步将这些值映射为黑色和白色.
I added the labels on top of each tile, rounded to one decimal place, with a % sign. In order to make them show up nicely against the dark background fill for high values, I made the color aesthetic for the labels a logical expression, then further down mapped those values to black and white.
然后无法分辨出 variable 的四种类型中的哪种类型，因此我通过将数据仅过滤到墨西哥和E并进行移位来标记右上角的那些类型.使用与以前相同的调整值，将y位置比图块的中心稍远一点.
Then there's no way to tell which of the four types of variable are in which tile, so I labeled the ones in the upper right corner by filtering the data to only Mexico and E, and shifting the y position a little bit further than the center of the tile using the same adjustment value as before.
然后我添加了一个分散的填充调色板，并在色条上标记了％符号.
Then I added a diverging fill palette and labeled the colourbar with % signs.
然后，您必须为x和y标度指定标签，因为它们现在只是数字位置.
Then you have to specify the labels for the x and y scales, since they're just numeric positions right now.
这篇关于将一个单元划分为一个单独的热图/具有多个颜色代码的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！