r - ggplot - 集中 facet_grid 标题并且只出现一次

Question

这个问题在这里已经有了答案:





Nested facets in ggplot2 spanning groups

(2 个回答)


去年关闭。




我在 ggplot 中创建了一个图表里面有两个变量 facet_grid .

我希望每个方面的标题仅在该方面的中心重复一次。

例如，第一个原始(上面)中的 0 和 1 将只出现一次并且出现在中间。

在我原来的情节中，每个方面的情节数量不相等。因此，使用 patchwork 将两个图拼凑在一起/cowplot/ggpubr效果不佳。
我更喜欢只使用 ggplot 的解决方案/hack .



样本数据:

df <- head(mtcars, 5)

示例图:

df %>% 
  ggplot(aes(gear, disp)) + 
  geom_bar(stat = "identity") + 
  facet_grid(~am + carb,
             space = "free_x", 
             scales = "free_x") +
  ggplot2::theme(
    panel.spacing.x = unit(0,"cm"), 
    axis.ticks.length=unit(.25, "cm"), 
    strip.placement = "outside",
    legend.position = "top",
    legend.justification = "center",
    legend.direction = "horizontal",
    legend.key.size = ggplot2::unit(1.5, "lines"),
    # switch off the rectangle around symbols
    legend.key = ggplot2::element_blank(),
    legend.key.width = grid::unit(2, "lines"),
    # # facet titles
    strip.background = ggplot2::element_rect(
      colour = "black",
      fill = "white"),
    panel.background = ggplot2::element_rect(
      colour = "white",
      fill = "white"), 
    panel.grid.major = element_blank(),
    panel.grid.minor = element_blank())

编辑 - 新数据

我创建了一个示例数据，它更准确地类似于我的实际数据。

structure(list(par = c("Par1", "Par1", "Par1", "Par1", "Par1", 
"Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", 
"Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", 
"Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par2", "Par2", 
"Par2"), channel_1 = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 11L, 11L, 11L, 11L, 
11L, 11L, 11L, 11L, 11L, 1L, 1L, 1L), .Label = c("Center", "Left \nFrontal", 
"Left \nFrontal Central", "Left \nCentral Parietal", "Left \nParietal Ooccipital", 
"Left", "Right \nFrontal", "Right \nFrontal Central", "Right \nCentral Parietal", 
"Right \nParietal Ooccipital", "Right"), class = "factor"), freq = structure(c(1L, 
1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 
3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L), .Label = c("Alpha", 
"Beta", "Gamma"), class = "factor"), group = c("a", "b", "c", 
"a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", 
"b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", 
"c"), m = c(0.488630500442935, 0.548666228768508, 0.0441536349332613, 
0.304475866391531, 0.330039488441422, 0.0980622573307064, 0.0963996979198171, 
0.301679466108907, 0.240618782227119, 0.35779695722622, 0.156116647839907, 
0.0274546218676152, 0.0752501569920047, 0.289342864254614, 0.770518960576786, 
0.548130676907356, 0.180158614358946, 0.238520826021687, 0.406326198917495, 
0.159739769132509, 0.140739952534666, 0.295427640977557, 0.106130817023844, 
0.214006898241167, 0.31081727835652, 0.366982521446529, 0.264432086988446, 
0.0761271112139142, 0.0811642772125171, 0.0700455890939194), 
    se = c(0.00919040825504951, 0.00664655073810519, 0.0095517721611042, 
    0.00657090455386036, 0.00451135146762504, 0.0188625074573698, 
    0.00875378313351897, 0.000569521129673224, 0.00691447732630984, 
    0.000241814142091401, 0.0124584589176995, 0.00366855139256551, 
    0.0072981677277562, 0.0160663614099261, 0.00359337442316408, 
    0.00919725279757502, 0.040856967817406, 0.00240910563984416, 
    0.0152236046767608, 0.00765487375180611, 0.00354140237391633, 
    0.00145468584619171, 0.0185141245423404, 0.000833307847848054, 
    0.0038193622895167, 0.0206130436440409, 0.0066911922721337, 
    7.3079999953491e-05, 0.0246233416039572, 0.00328150956514463
    )), row.names = c(NA, -30L), class = c("tbl_df", "tbl", "data.frame"
))

阴谋:

df %>%
  ggplot(aes(channel_1, m, 
             group = group,
             fill = group, 
             color = group)) +
  facet_grid(~par + freq,
             space="free_x", 
             scales="free_x") +
  geom_errorbar(
    aes(min = m - se, ymax = m + se, alpha = 0.01), 
    width = 0.2, size = 2, color = "black", 
    position = position_dodge(width = 0.6)) +
  geom_bar(stat = "identity",
           position = position_dodge(width = 0.6),
           # color = "black", 
           # fill = "white",
           width = 0.6, 
           size = 2, aes(alpha = 0.01))  + 
  scale_shape_manual(values = c(1, 8, 5)) + 
  labs(
    color = "",
    fill = "", 
    shape = "") +
  guides(
    color = FALSE,
    shape = FALSE) +
  scale_alpha(guide = "none")

Answer 1

最快的黑客:用情节伪造方面并结合。这需要一些黑客攻击，但它可能仍然比搞乱 grobs 更少黑客攻击:

为刻面图制作联合变量。

制作假面并结合拼凑等包装。将地块的边距减少到负，这样就真的没有边距了。

使相对高度比有点高得离谱，所以第二个图消失了，只剩下刻面条。

library(patchwork)
library(tidyverse)

df <- head(mtcars,5)
df <- df %>% mutate(am_carb = factor(paste(am,carb,sep = '_'), 
                      labels = c( ' 1','2','1','4')))
##note!! the blank space in ' 1' label is on purpose!!! this is to make those labels unique, otherwise it would consider both '1' the same category!!

p1 <-
  df %>% 
  ggplot(aes(gear, disp)) + 
  geom_bar(stat = "identity") + 
  facet_grid(~am_carb, scales = "free_x") +
theme(panel.spacing.x = unit(0,"cm"),
      plot.margin = margin(t = -2),
      strip.background = element_rect(colour = "black",fill = "white"),
      panel.background = element_rect(colour = "white", fill = "white"), 
      panel.grid.major = element_blank(),
      panel.grid.minor = element_blank())

p2 <-
  df %>% 
  ggplot(aes(gear, disp)) + 
  geom_blank() + 
  facet_grid(~ am, scales = "free_x") +
  theme(panel.spacing.x = unit(0,"cm"),
        axis.text = element_blank(),
        axis.ticks = element_blank(),
        axis.title = element_blank(),
        plot.margin = margin(b = -2),
        strip.background = element_rect(colour = "black",fill = "white"),
        panel.background = element_rect(colour = "white", fill = "white"), 
        panel.grid.major = element_blank(),
        panel.grid.minor = element_blank())


p2/p1 + plot_layout(heights = c(0.1,100) )

创建于 2020-03-24 由 reprex package (v0.3.0)

用新数据更新 - 一些更复杂的方面。事实上，拼凑在这里很困难。在将假面转换为网格对象并更改宽度后，更容易将假面与牛图结合起来。内所有 cowplot .

mydat <- structure(list(par = c("Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par2", "Par2", "Par2"), channel_1 = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 11L, 11L, 11L, 11L, 11L, 11L, 11L, 11L, 11L, 1L, 1L, 1L), .Label = c("Center", "Left \nFrontal", "Left \nFrontal Central", "Left \nCentral Parietal", "Left \nParietal Ooccipital", "Left", "Right \nFrontal", "Right \nFrontal Central", "Right \nCentral Parietal", "Right \nParietal Ooccipital", "Right"), class = "factor"), freq = structure(c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L), .Label = c("Alpha", "Beta", "Gamma"), class = "factor"), group = c("a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c"), m = c(0.488630500442935, 0.548666228768508, 0.0441536349332613, 0.304475866391531, 0.330039488441422, 0.0980622573307064, 0.0963996979198171, 0.301679466108907, 0.240618782227119, 0.35779695722622, 0.156116647839907, 0.0274546218676152, 0.0752501569920047, 0.289342864254614, 0.770518960576786, 0.548130676907356, 0.180158614358946, 0.238520826021687, 0.406326198917495, 0.159739769132509, 0.140739952534666, 0.295427640977557, 0.106130817023844, 0.214006898241167, 0.31081727835652, 0.366982521446529, 0.264432086988446, 0.0761271112139142, 0.0811642772125171, 0.0700455890939194), se = c(0.00919040825504951, 0.00664655073810519, 0.0095517721611042, 0.00657090455386036, 0.00451135146762504, 0.0188625074573698, 0.00875378313351897, 0.000569521129673224, 0.00691447732630984, 0.000241814142091401, 0.0124584589176995, 0.00366855139256551, 0.0072981677277562, 0.0160663614099261, 0.00359337442316408, 0.00919725279757502, 0.040856967817406, 0.00240910563984416, 0.0152236046767608, 0.00765487375180611, 0.00354140237391633, 0.00145468584619171, 0.0185141245423404, 0.000833307847848054, 0.0038193622895167, 0.0206130436440409, 0.0066911922721337, 7.3079999953491e-05, 0.0246233416039572, 0.00328150956514463)), row.names = c(NA, -30L), class = c("tbl_df", "tbl", "data.frame"))

library(tidyverse)
library(cowplot)
#> 
#> ********************************************************
#> Note: As of version 1.0.0, cowplot does not change the
#>   default ggplot2 theme anymore. To recover the previous
#>   behavior, execute:
#>   theme_set(theme_cowplot())
#> ********************************************************

mydat <- mydat %>% mutate(par_freq = factor(paste(par,freq,sep = '_'), labels = c('Alpha', 'Beta', 'Gamma', 'Gamma ' )))

p1 <-
  mydat %>% 
    ggplot(aes(channel_1, m, group = group, fill = group, color = group)) +
  geom_bar(stat = "identity") + 
  facet_grid( ~ par_freq, scales = "free_x", space="free_x") +
  theme(panel.spacing.x = unit(0,"cm"),
        plot.margin = margin(t = -2),
        strip.background = element_rect(colour = "black",fill = "white"),
        panel.background = element_rect(colour = "white", fill = "white"), 
        panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        legend.position = 'none')

p2 <-
  mydat %>%
    ggplot(aes(channel_1, m, group = group, fill = group, color = group)) +
    geom_blank() + 
    facet_grid(~ par) +
    theme(panel.spacing.x = unit(0,"cm"),
          axis.text = element_blank(),
          axis.ticks = element_blank(),
          axis.title = element_blank(),
          plot.margin = margin(b = -2),
          strip.background = element_rect(colour = "black",fill = "white"),
          panel.background = element_rect(colour = "white", fill = "white"), 
          panel.grid.major = element_blank(),
          panel.grid.minor = element_blank())


gt <-  cowplot::as_gtable(p2)
gt$widths[5] <- 8*gt$widths[7]
cowplot::plot_grid(gt, p1, align = "v", axis = 'l',nrow = 2, rel_heights = c(5, 100))
# you need to play around with the values unfortunately. 

创建于 2020-03-24 由 reprex package (v0.3.0)

一些额外的想法

我在想一个人无法绕过这样的黑客 - 因为原始图的 gtable_layout(带有两个方面变量)显示整个方面条是一个 grob! THis answer proved me wrong - the grob contains a nested table for both strips! .但由于 ggnomics，有一个更简单的解决方案。包 - 见我的第二个答案

p_demo <-   ggplot(mydat, aes(channel_1, m)) + 
  geom_bar(stat = "identity") + 
  facet_grid(~par +freq , space = "free_x",  scales = "free_x") +
  theme(panel.spacing.x = unit(0,"cm"))

gt <-  cowplot::as_gtable(p_demo)
gtable::gtable_show_layout(gt)

创建于 2020-03-24 由 reprex package (v0.3.0)