r - 更改绘制线的顺序，以便在 ggplot 中的时间序列中位于顶部

Question

所以我正在为我正在从事的项目在 ggplot 上绘制一些时间序列数据。数据如下所示： 示例数据：

structure(list(Date = c("2015-01-01", "2015-02-01", "2015-03-01", 
"2015-04-01"), Actual = c(500L, 600L, 700L, 750L), Fcst1 = c(600L, 
610L, 634L, 650L), Fcst2 = c(500L, 630L, 875L, 900L), Fcst3 = c(500L, 
600L, 754L, 800L), Fcst4 = c(500L, 600L, 700L, 760L)), class = "data.frame", row.names = c(NA, 
-4L))

Date        Actual Fcst1 Fcst2 Fcst3 Fcst4
2015-01-01  500    600   500   500   500
2015-02-01  600    610   630   600   600
2015-03-01  700    634   875   754   700
2015-04-01  750    650   900   800   760
..........  ...    ...   ...   ...   ...

数据本身又持续了 60 个月，总共有 40 个预测，每月调整一次。当我尝试绘制它时，我希望实际线位于顶部，但它最终被首先绘制。这是我正在使用的代码：

df <- df %>%
  mutate(Date = ymd(Date))
colnames(df)[3:length(df)] <-
  paste("df", colnames(df[, c(3:length(df))]), sep = "")
ggplot(
  tidyr::pivot_longer(df, c(Fcst1:Fcst6, Actual), names_to = "Forecast", names_prefix = "df"),
  aes(Date, value, color = Forecast)
) +
  geom_line(size = 1) +
  scale_color_manual(values = c(
    "Fcst1" = "red", "Fcst2" = "blue",
    "Fcst3" = "green", "Fcst4" = "yellow", "Fcst5" = "purple",
    "Fcst6" = "orange", "Actual" = "black"
  )) +
  ggtitle(label = "Actuals vs Forecasts", subtitle = "Dataset") +
  ylab("Rate") +
  scale_y_continuous(labels = scales::comma)

我想保留图例中的顺序，所以我希望 Actuals 在顶部，然后是scale_color_manual. 现在，实际值首先被绘制（这意味着它在所有其他预测之下），我希望它在顶部（最好用更粗的线，也许size=1.2。谢谢！

标签： rggplot2dplyrtidyverse

Answer 1

这是使用您包含的数据的可能解决方案。您必须对图例的级别进行格式化Forecast和修改scale_color_manual()。我为此添加了一个技巧：

library(tidyverse)
#Data
df <- structure(list(Date = c("2015-01-01", "2015-02-01", "2015-03-01", 
"2015-04-01"), Actual = c(500L, 600L, 700L, 750L), Fcst1 = c(600L, 
610L, 634L, 650L), Fcst2 = c(500L, 630L, 875L, 900L), Fcst3 = c(500L, 
600L, 754L, 800L), Fcst4 = c(500L, 600L, 700L, 760L)), class = "data.frame", row.names = c(NA, 
-4L))

#Format date
df <- df %>% mutate(Date = ymd(Date))
#Create data for plot
df2 <- tidyr::pivot_longer(df, c(Fcst1:Fcst4, Actual), names_to = "Forecast", names_prefix = "df")
#Format levels 
labs <- unique(df2$Forecast)
i1 <- labs[which(labs=='Actual')]
i2 <- rev(labs[which(labs!='Actual')])
i3 <- c(i2,i1)
df2$Forecast <- factor(df2$Forecast,levels=i3,ordered = T)
#Plot
ggplot(df2,aes(Date, value, color = Forecast)) +
  geom_line(size = 1) +
  scale_color_manual(values = c(
    "Fcst1" = "red", "Fcst2" = "blue",
    "Fcst3" = "green", "Fcst4" = "yellow", "Fcst5" = "purple",
    "Fcst6" = "orange", "Actual" = "black"
  ),guide = guide_legend(reverse=TRUE)) +
  ggtitle(label = "Actuals vs Forecasts", subtitle = "Dataset") +
  ylab("Rate") +
  scale_y_continuous(labels = scales::comma)

输出：