根据条件更改` `ggplot`‘标签颜色

Question

我有一个关于根据条件改变geom_text & geom_label的颜色的问题。

预期输出

• 数据标签为2021年的color = blue，color = grey为2020年的
• 缺口标签为negative时为红色，当为color = grey时为绿色

数据

data <-  tibble(Factor = c("A", "B", "A", "B"),
   Score = c(90, 85, 85, 90),
   Year = c("2020", "2020", "2021", "2021"))

数据转换

   df_2 = data %>% 
     pivot_wider(names_from = Year, names_prefix = "Year", values_from = Score) %>% 
     mutate(gap = Year2021 - Year2020) %>% 
     select(Factor, gap)

图

data %>% 
  left_join(df_2) %>% 
  ggplot(aes(x = Factor, y = Score, fill = Year)) +
  geom_col(position='dodge') +
  geom_text(aes(label=Score), 
            position=position_dodge(width = 0.9), 
            vjust=-0.40) +
  geom_label(aes(y = 100, label = ifelse(Year == 2021, gap, NA_character_)), na.rm = TRUE) +
  scale_y_continuous(limits = c(0,105))
  theme_minimal()

​

​

预期输出

• 85 & 90 = blue为2021年，90 & 85 = grey为2020
• -5 = red和5 = green

Answer 1

第二个选项是对文本和条形颜色使用手动缩放。对于标签，我们可以使用ggnewscale包，它允许为相同的美学多个尺度。这样，我们可以添加第二个填充比例：

library(ggplot2)
library(ggnewscale)

ggplot(dat, aes(x = Factor, y = Score, fill = Year)) +
  geom_col(position = "dodge") +
  geom_text(aes(label = Score, color = Year),
    position = position_dodge(width = 0.9),
    vjust = -0.40
  ) +
  scale_color_manual(aesthetics = c("fill", "color"), values = c("2021" = "blue", "2020" = "grey")) +
  ggnewscale::new_scale_fill() +
  geom_label(aes(y = 100, label = ifelse(Year == 2021, gap, NA_character_), fill = gap > 0), na.rm = TRUE) +
  scale_fill_manual(values = c("FALSE" = "red", "TRUE" = "green")) +
  scale_y_continuous(limits = c(0, 105)) +
  theme_minimal()

​

​

数据

data <- data.frame(
  Factor = c("A", "B", "A", "B"),
  Score = c(90, 85, 85, 90),
  Year = c("2020", "2020", "2021", "2021")
)

library(dplyr)
library(tidyr)

df_2 <- data %>%
  pivot_wider(names_from = Year, names_prefix = "Year", values_from = Score) %>%
  mutate(gap = Year2021 - Year2020) %>%
  select(Factor, gap)

dat <- data %>%
  left_join(df_2)