ggplot2：有條件地改變線型和顏色

Question

我有以下數據框：

df <- tribble(
  ~group,             ~year,   ~value,  
  "group1",     "2010-2015",       70,
  "group2",     "2010-2015",       75,
  "group1",     "2015-2020",       72,
  "group2",     "2015-2020",       77,
  "group1",     "2020-2025",       74,
  "group2",     "2020-2025",       79,
  "group1",     "2025-2030",       76,
  "group2",     "2025-2030",       81)

我想繪制為x 、年份和y值，按組着色，即，

ggplot(df, aes(year, value, color = group, group = group)) +
  geom_line(size = 1.1)

但我希望“2020-2025”和“2025-2030”年是虛線。

使用以下代碼，我收到此錯誤：“錯誤：geom_path：如果您使用點線或虛線，顏色、大小和線型必須在線上保持不變”

df %>% 
  mutate(projection = year %in% c("2020-2025", "2025-2030")) %>% 
  ggplot(aes(year, value, color = group, group = group)) +
  geom_line(aes(linetype = projection), size = 1.1)

Answer 1

試試這個

library(tidyr)
library(dplyr)
library(ggplot2)

df %>% 
  mutate(projection = list("2010-2015" = F, "2015-2020" = c(T, F), "2020-2025" = T, "2025-2030" = T)[year]) %>% 
  unnest(projection) %>% 
  ggplot(aes(year, value, color = group, group = interaction(group, projection))) +
  geom_line(aes(linetype = projection), size = 1.1)

輸出

---

另一種形式

df %>% 
  mutate(projection = list("2010-2015" = F, "2015-2020" = F, "2020-2025" = c(T, F), "2025-2030" = T)[year]) %>% 
  unnest(projection) %>% 
  ggplot(aes(year, value, color = group, group = interaction(group, projection))) +
  geom_line(aes(linetype = projection), size = 1.1)

輸出

Answer 2

這可以通過兩個geom_line圖層並過濾每個圖層要顯示的數據來實現：

library(tidyverse)

df <- tribble(
  ~group,             ~year,   ~value,  
  "group1",     "2010-2015",       70,
  "group2",     "2010-2015",       75,
  "group1",     "2015-2020",       72,
  "group2",     "2015-2020",       77,
  "group1",     "2020-2025",       74,
  "group2",     "2020-2025",       79,
  "group1",     "2025-2030",       76,
  "group2",     "2025-2030",       81)

df1 <- df %>% 
  mutate(projection = year %in% c("2020-2025", "2025-2030"))

ggplot(mapping = aes(year, value, color = group, group = group)) +
  geom_line(data = filter(df1, !year %in% c("2025-2030")), size = 1.1) +
  geom_line(data = filter(df1, projection), size = 1.1, linetype = "dashed")