Order legend according to the color sequence in geom

I am creating a stack plot for a data of 3 years, using geom_area. It is working perfectly fine, except that I would like the legend order as per the color sequence. I am mentioning the sample data and code below:

#sample dataset
df <- structure(list(
  Topic = c("Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_I", "Topic_I", "Topic_I", "Topic_I", "Topic_I", "Topic_I", "Topic_I"), Frequency = c(22L, 27L, 68L, 34L, 60L, 13L, 23L, 11L, 12L, 23L, 19L, 22L, 9L, 36L, 127L, 83L, 37L, 19L, 31L, 19L, 32L, 25L, 20L, 17L, 19L, 22L, 48L, 32L, 36L, 71L, 37L, 12L, 10L, 37L, 48L, 11L, 15L, 10L, 9L, 19L, 19L, 14L, 22L, 51L, 14L, 33L, 22L, 36L, 16L, 10L, 31L, 36L, 15L, 10L, 20L, 23L, 21L, 14L, 20L, 17L, 18L, 22L, 9L, 14L, 37L, 18L, 14L, 15L, 17L, 47L, 28L, 34L, 68L, 227L, 22L, 25L, 8L, 11L, 5L, 2L, 4L, 34L, 8L, 21L, 4L, 5L, 33L, 19L, 12L, 18L, 22L, 23L, 38L, 96L, 112L, 53L, 87L, 103L, 58L, 44L),
  Timestamp = structure(c(17897, 17928, 17957, 17988, 18018, 18048, 18079, 18109, 18169, 18200, 18230, 18233, 18262, 18302, 18323, 18353, 18383, 18414, 18444, 18475, 18535, 18566, 18596, 18599, 18628, 18660, 18713, 18744, 18770, 18799, 18827, 18856, 18884, 18913, 18944, 18628, 18660, 18713, 18744, 18770, 18799, 18827, 18856, 18884, 18913, 18944, 18262, 18302, 18323, 18353, 18383, 18414, 18444, 18475, 18535, 18566, 18596, 18599, 18628, 18660, 18713, 18744, 18770, 18799, 18827, 18856, 18884, 18913, 18944, 17897, 17928, 17957, 17988, 18018, 18048, 18079, 18109, 18169, 18200, 18230, 18233, 18262, 18302, 18323, 18353, 18383, 18414, 18444, 18475, 18535, 18566, 18596, 18599, 17897, 17928, 17957, 17988, 18018, 18048, 18079), class = "Date")
), row.names = c(5L, 10L, 15L, 20L, 25L, 30L, 35L, 40L, 45L, 50L, 55L, 60L, 61L, 66L, 71L, 76L, 81L, 86L, 91L, 96L, 101L, 106L, 111L, 116L, 122L, 127L, 132L, 137L, 142L, 147L, 152L, 157L, 162L, 167L, 172L, 124L, 129L, 134L, 139L, 144L, 149L, 154L, 159L, 164L, 169L, 174L, 62L, 67L, 72L, 77L, 82L, 87L, 92L, 97L, 102L, 107L, 112L, 117L, 125L, 130L, 135L, 140L, 145L, 150L, 155L, 160L, 165L, 170L, 175L, 3L, 8L, 13L, 18L, 23L, 28L, 33L, 38L, 43L, 48L, 53L, 58L, 63L, 68L, 73L, 78L, 83L, 88L, 93L, 98L, 103L, 108L, 113L, 118L, 1L, 6L, 11L, 16L, 21L, 26L, 31L), class = "data.frame")

Below is the simple code I am using to produce the plot:

ggplot(df, aes(x=Timestamp, y=Frequency, fill=Topic))   
  scale_x_date(date_breaks = '1 month', date_labels = "%b-%y") 
  geom_area(alpha=0.6 , size=1, colour="black", position = position_fill()) 
  scale_fill_discrete(guide = guide_legend())   #, breaks = df_q$Topic #to sort legends
  theme(legend.position="bottom", legend.box = "horizontal") 
  ggtitle("Questionable") 
  xlab("Month") 
  guides(fill = guide_legend(nrow = 1))

I am getting the following output:

Is it possible to order the legend (ignoring Topic order) as per the sequence of color from left(2019) to right(2021) and top to bottom? For instance, orange (top color in 2019) should be first one in legend, blue second and so on. And it would be really great to know if there's a way each legend just below its corresponding year, to easily differntiate the topic for each year.

CodePudding user response：

Just specify the order manually with the limits argument to scale_fill

library(ggplot2)

df <- structure(list(
  Topic = c("Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_A", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_B", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_C", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_D", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_E", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_F", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_G", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_H", "Topic_I", "Topic_I", "Topic_I", "Topic_I", "Topic_I", "Topic_I", "Topic_I"), Frequency = c(22L, 27L, 68L, 34L, 60L, 13L, 23L, 11L, 12L, 23L, 19L, 22L, 9L, 36L, 127L, 83L, 37L, 19L, 31L, 19L, 32L, 25L, 20L, 17L, 19L, 22L, 48L, 32L, 36L, 71L, 37L, 12L, 10L, 37L, 48L, 11L, 15L, 10L, 9L, 19L, 19L, 14L, 22L, 51L, 14L, 33L, 22L, 36L, 16L, 10L, 31L, 36L, 15L, 10L, 20L, 23L, 21L, 14L, 20L, 17L, 18L, 22L, 9L, 14L, 37L, 18L, 14L, 15L, 17L, 47L, 28L, 34L, 68L, 227L, 22L, 25L, 8L, 11L, 5L, 2L, 4L, 34L, 8L, 21L, 4L, 5L, 33L, 19L, 12L, 18L, 22L, 23L, 38L, 96L, 112L, 53L, 87L, 103L, 58L, 44L),
  Timestamp = structure(c(17897, 17928, 17957, 17988, 18018, 18048, 18079, 18109, 18169, 18200, 18230, 18233, 18262, 18302, 18323, 18353, 18383, 18414, 18444, 18475, 18535, 18566, 18596, 18599, 18628, 18660, 18713, 18744, 18770, 18799, 18827, 18856, 18884, 18913, 18944, 18628, 18660, 18713, 18744, 18770, 18799, 18827, 18856, 18884, 18913, 18944, 18262, 18302, 18323, 18353, 18383, 18414, 18444, 18475, 18535, 18566, 18596, 18599, 18628, 18660, 18713, 18744, 18770, 18799, 18827, 18856, 18884, 18913, 18944, 17897, 17928, 17957, 17988, 18018, 18048, 18079, 18109, 18169, 18200, 18230, 18233, 18262, 18302, 18323, 18353, 18383, 18414, 18444, 18475, 18535, 18566, 18596, 18599, 17897, 17928, 17957, 17988, 18018, 18048, 18079), class = "Date")
), row.names = c(5L, 10L, 15L, 20L, 25L, 30L, 35L, 40L, 45L, 50L, 55L, 60L, 61L, 66L, 71L, 76L, 81L, 86L, 91L, 96L, 101L, 106L, 111L, 116L, 122L, 127L, 132L, 137L, 142L, 147L, 152L, 157L, 162L, 167L, 172L, 124L, 129L, 134L, 139L, 144L, 149L, 154L, 159L, 164L, 169L, 174L, 62L, 67L, 72L, 77L, 82L, 87L, 92L, 97L, 102L, 107L, 112L, 117L, 125L, 130L, 135L, 140L, 145L, 150L, 155L, 160L, 165L, 170L, 175L, 3L, 8L, 13L, 18L, 23L, 28L, 33L, 38L, 43L, 48L, 53L, 58L, 63L, 68L, 73L, 78L, 83L, 88L, 93L, 98L, 103L, 108L, 113L, 118L, 1L, 6L, 11L, 16L, 21L, 26L, 31L), class = "data.frame")

ggplot(df, aes(x=Timestamp, y=Frequency, fill=Topic))   
  scale_x_date(date_breaks = '1 month', date_labels = "%b-%y") 
  geom_area(alpha=0.6 , size=1, colour="black", position = position_fill()) 
  theme(legend.position="bottom", legend.box = "horizontal") 
  guides(fill = guide_legend(nrow = 1))  
  scale_fill_discrete(limits = paste("Topic", LETTERS[c(1,7,9,2,5,8,3,4,6)], sep = "_"))

In order to get a separate legend per year, you basically need to create separate plots with respective legends. The main challenge is to assign the correct color values and guide limits.

library(tidyverse)
library(patchwork)

## first get the colors
my_cols <- scales::hue_pal()(9)
## assign them to the order you like
names(my_cols) = paste("Topic", LETTERS[c(1,7,9,2,5,8,3,4,6)], sep = "_")

p_list <- 
  df %>%
  ## create year variable by which you split into a list
  mutate(year = lubridate::year(Timestamp)) %>%
  split(.$year) %>%
  ## pass this list to a loop function to create three separate plots 
  map(~ggplot(data = .x, aes(x=Timestamp, y=Frequency, fill=Topic))   
  scale_x_date(date_breaks = '1 month', date_labels = "%b-%y") 
  geom_area(alpha=0.6 , size=1, colour="black", position = position_fill()) 
  theme(legend.position="bottom", legend.box = "horizontal") 
    ## changing label position because of crowding
  guides(fill = guide_legend(nrow = 1, label.position = "bottom"))  
  ## you will need to set the limits to the unique values in each plot
    ## I am also removing the guide title because of the visual crowding
  scale_fill_manual(NULL, values = my_cols, limits = unique(.x$Topic))
  
  )

## patchwork is great for such things
wrap_plots(p_list)