ggplot：显示具有多层的自定义图例

Question

我有一个带有几层条形图和点的 ggplot。我不知道如何正确显示每一层的图例。这是代码：

country <- c(rep(c("AA"), 4), rep(c("BB"), 4))
score <- c(1.3, 7.2, 4.9, 7.5, 6.8, 4.1, 4.9, 5.8)
y_axis <- c(rep(c("Score 2019", "Score 2020", "Group median", "Cluster average"), 2))

dt <- data.table(country, score, y_axis)


ggplot(, aes(country, score)) +
  theme_bw() +
  geom_bar(data=dt[y_axis=="Score 2019"], stat="identity", fill="grey", width=0.5) +
  geom_point(data=dt[y_axis=="Score 2020"], aes(shape="Score 2020"), color="green4", size=2) +
  geom_point(data=dt[y_axis=="Cluster average"], aes(shape="Cluster average"), color="orange", size=2) +
  geom_hline(yintercept=unique(dt[y_axis=="Group median"][,score]), size=1, color="red3")

现在我只能在图例中显示“Cluster average”和“Score 2020”（颜色错误）。如何正确显示图例中的所有图层（分数 2019 和 2020，中位数和平均值）？

Answer 1

为了正确显示点的图例，您需要将颜色和形状映射到美学。将 hline 分开并将其映射到不同的 non-consequential 美学，例如 linetype 以使其出现在图例中。

ggplot(, aes(country, score)) +
  theme_bw(base_size = 20) +
  geom_col(data = dt[y_axis == "Score 2019"], 
           aes(fill = "Score"), width = 0.5) +
  geom_point(data = dt[y_axis == "Score 2020"], 
             aes(shape = "Score 2020", color = "Score 2020"), size = 4) +
  geom_point(data = dt[y_axis == "Cluster average"], size = 4,
             aes(shape="Cluster average", color = "Cluster average")) +
  geom_hline(data =  dt[y_axis == "Group median"],
             aes(yintercept = score, linetype = "Group median"), 
             color = "red4", size = 1) +
  scale_shape_manual(values = c("Cluster average" = 16,
                                "Score 2020" = 17), name = "Legend") +
  scale_color_manual(values = c("Cluster average" = 'orange', 
                                "Score 2020" = "green4"), name = "Legend") +
  scale_fill_manual(values = "gray") +
  labs(linetype = "") +
  theme(legend.title = element_blank(),
        legend.spacing.y = unit(0, "mm"),
        legend.margin = margin(0, 0, 0, 0))

ggplot：显示具有多层的自定义图例

ggplot: showing custom legend with multiple layers

r

ggplot2