R-来自不同数据集但相同颜色传说的盒子图

Question

我有两个表示相同的数据集，但一个来自模拟数据，另一个来自真实数据。我想将这两个与盒子图进行比较。到目前为止，我确实像您在图像中看到的那样绘制它们。问题是，我希望一个组中的每个盒子图都具有不同的颜色，但对于两个数据集都相同，因此，模拟数据的alpha = 0.2，只有5种不同的颜色。假设，要将真实的“ dt”'rfr'与模拟的“ dt”“ rfr”进行比较，我想拥有两个带有相同颜色的盒子图，但一个“ alphaed'。

此外，我不知道该如何在两列传奇中展示，一个用于真实标签，一个用于模拟标签（在开始的标签上带有's'的标签）。

我的绘图代码是以下

p <- ggplot()+geom_boxplot(data=simulation,aes(x=param,y=data, fill=algo), alpha=0.2)+
  geom_boxplot(data=ADCF2param_shuffle,aes(x=param, y = data, fill=algo))+
  geom_point(data =gt_vs_fitted,aes(param, y = data), color='red', size=4, shape=18)
p+scale_fill_brewer(palette="YlGnBu") + theme_classic() + labs(y="CCC", x= "Parameters")

在这里一些示例数据。对于每个“算法”，我都有每个“参数”的数据。以及模拟数据中相同的结构。

Answer 1

尝试以下操作：

library(tidyverse)

# Made-up data
simulation <- tribble(
  ~param, ~data, ~algo,
  "Dt", 1, "GBR",
  "Dt", 1.3, "GBR",
  "Dt", 1.5, "ETR",
  "Dt", 1.7, "ETR",
  "Dv", 1.5, "ETR",
  "Dv", 1.6, "ETR",
  "Dv", 2, "GBR",
  "Dv", 2.1, "GBR"
)

shuffle <- tribble(
  ~param, ~data, ~algo,
  "Dt", 2.1, "GBR",
  "Dt", 2.2, "GBR",
  "Dt", 2.4, "ETR",
  "Dt", 2.6, "ETR",
  "Dv", 2.4, "ETR",
  "Dv", 2.8, "ETR",
  "Dv", 3.2, "GBR",
  "Dv", 2.9, "GBR"
)

fitted <- tribble(
  ~param, ~data, ~algo,
  "Dt", 1.12, "GBR",
  "Dt", 1.54, "ETR",
  "Dv", 1.56, "ETR",
  "Dv", 2.12, "GBR"
)

joined_df <- bind_rows(
  simulation = simulation,
  shuffle = shuffle,
  .id = "dataset"
)

# Plot
joined_df |> 
  ggplot(aes(param, data, fill = algo)) +
  geom_boxplot(aes(alpha = dataset, linetype = dataset), 
               data = joined_df |> filter(dataset == "simulation")) + 
  geom_boxplot(aes(alpha = dataset, linetype = dataset), 
               data = joined_df |> filter(dataset == "shuffle")) + 
  geom_point(data = fitted, color = "red", size = 2, shape = 18) +
  scale_fill_brewer(palette = "YlGnBu") + 
  scale_alpha_manual(values = c(1, 0.2)) +
  theme_classic() + 
  labs(y = "CCC", x = "Parameters") +
  theme(legend.position = "bottom")

^{在2022-07-04创建的 reprex软件包 （v2.0.1）}