如何使用facet_grid创建翻转分组小提琴图，并在R中使用ggplot2保持相同比例的小提琴图

Question

我想创建一个翻转的小提琴图（在 geom_point() 中为组平均值添加额外点，在 geom_segment() 中添加从最小值到最大值的线）。我使用facet_grid()将这些组集成为facet。 我唯一无法编辑的是每个方面的小提琴图的宽度。所有组都具有相同数量的值，并且应在小提琴图中以相同的小提琴图面积表示。当不使用构面时，我在绘制点和线段时遇到麻烦。 您知道如何实现这一目标吗？

输入数据：

dat1 <- data.frame(subgroup = c(rep(("mountain"), 150), rep(("upland"), 150), rep(("riparian"), 150)),
                        group = rep(c("birds"), 450),
                        value = sample(c(1:100), 450, replace = TRUE))
dat2 <- data.frame(subgroup = c(rep(("mountain"), 150), rep(("upland"), 150), rep(("riparian"), 150)),
                   group = rep(c("beetles"), 450),
                   value = sample(c(1:50), 450, replace = TRUE))
plot_data <- rbind(dat1, dat2)
plot_data$subgroup <- factor(plot_data$subgroup, levels = c("mountain", "upland","riparian") )

到目前为止我的情节代码：

library(ggplot2)
library(dplyr)

plot_data %>%
  ggplot( aes(x = interaction(group, subgroup), y = value, fill = subgroup, label = group)) +
  geom_violin(width = 1, alpha = 0.5, scale = "area")+
  geom_segment(data = plot_data %>%
                 dplyr::group_by(group, subgroup) %>%
                 dplyr::summarise(lower = min(value, na.rm =T), 
                                  upper = max(value, na.rm = T)),
               aes(x = interaction(group, subgroup) , xend = interaction(group, subgroup),
                   y = lower, yend = upper), linewidth = 1, col = "grey10", alpha = 0.8)+
  geom_point(data = plot_data %>%
             dplyr::group_by(group, subgroup) %>%
               dplyr::summarise(mean = mean(value, na.rm = T)), 
             aes(x= interaction(group, subgroup), y = mean), shape = 21, size = 4, col = "black")+
  coord_flip()+
  labs(y= "Model values", x="")+
  facet_grid(group ~ ., scales = "free_y")+
  theme_minimal()+
  scale_x_discrete(labels= rev(c("riparian",  "upland", "mountain")))+
  theme(
    axis.text.y = element_text(size = 12), 
    axis.text.x = element_text(size = 10),
    axis.title.y = element_text(size = 20), 
    axis.title.x = element_text(size = 12),
    strip.text.y = element_text(size = 14),  # Customize facet labels
    legend.position = "none")

绘图如下所示：

我遇到的主要问题是小提琴宽度，这里显示正确：

plot_data %>% 
  ggplot(aes(y = group, x = value, fill = subgroup)) +
  geom_violin(width = 1, alpha = 0.5, scale = "area")

但在这里我很难为子组添加点、线和额外标签而不是图例（如上所示）

Answer 1

您的宽度不正确，因为您使用具有自由缩放比例的面。

要重新排序数据，您可以更改

interaction

中变量的顺序。十你不需要使用方面。您可以添加组标签作为文本注释：

plot_data %>%
  ggplot( aes(y = interaction(subgroup, group), x = value, fill = subgroup, label = group)) +
  geom_violin(width = 1, alpha = 0.5, scale = "area", orientation = "y") +
  geom_segment(data = plot_data %>%
                 dplyr::group_by(group, subgroup) %>%
                 dplyr::summarise(lower = min(value, na.rm =T), 
                                  upper = max(value, na.rm = T)),
               aes(y = interaction(subgroup, group) , yend = interaction(subgroup, group),
                   x = lower, xend = upper), linewidth = 1, col = "grey10", alpha = 0.8) +
  geom_point(data = plot_data %>%
               dplyr::group_by(subgroup, group) %>%
               dplyr::summarise(mean = mean(value, na.rm = T)), 
             aes(y = interaction(subgroup, group), x = mean), shape = 21, size = 4, col = "black") +
  # coord_flip() +
  labs(x = "Model values", y = "") +
  annotate(geom = "text", label = "birds", x = 100, y = 5, angle = 90, vjust = 1) +
  annotate(geom = "text", label = "beetles", x = 100, y = 2, angle = 90, vjust = 1) +
  # facet_grid(group ~ ., scales = "free_y")+
  theme_minimal()+
  scale_y_discrete(labels= rev(rep(c("riparian",  "upland", "mountain"), 2)))+
  theme(
    axis.text.y = element_text(size = 12), 
    axis.text.x = element_text(size = 10),
    axis.title.y = element_text(size = 20), 
    axis.title.x = element_text(size = 12),
    strip.text.y = element_text(size = 14),  # Customize facet labels
    legend.position = "none")