Manually adding legend to ggplot-CodePudding

I made a graph of some GPS-data, where the colour of the path indicates in which habitat a bird was. I added a column to the df, with the desired colours as I wanted to set those manually per habitat type. This is the code I used (which works):

worldmap <- ne_countries(scale = 'medium', type = 'map_units',
                         returnclass = 'sf')
xlim <- c(min(random_bird$location.long)-0.1, max(random_bird$location.long) 0.1)
ylim <- c(min(random_bird$location.lat)-0.1, max(random_bird$location.lat) 0.1)
ggplot()   geom_sf(data = worldmap)  
  coord_sf(xlim = xlim, ylim = ylim, expand = FALSE)  
  theme_bw() 
  geom_path(aes(x = random_bird$location.long, y = random_bird$location.lat), colour = (random_bird$colour))  
  labs(x="Longitude", y="Latitude")   ggtitle(random_bird$colour_ring)  theme(plot.title = element_text(hjust = 0.5))

This is a snip of the dataframe:

                timestamp location.long location.lat        habitat   colour colour_ring
 2022-08-26 14:34:33.000      2.732021     51.15316 Not classified    white       S.ROP
 2022-08-26 19:34:30.000      2.735457     51.15966     Intertidal cornsilk       S.ROP
 2022-08-27 03:34:31.000      2.741388     51.16101     Intertidal cornsilk       S.ROP
 2022-08-27 05:34:30.000      2.736342     51.15954     Intertidal cornsilk       S.ROP
 2022-08-27 08:34:56.000      2.728476     51.15704     Intertidal cornsilk       S.ROP
 2022-08-27 10:34:28.000      2.726892     51.15534     Intertidal cornsilk       S.ROP
 2022-08-27 11:34:28.000      2.726382     51.15453 Not classified    white       S.ROP
 2022-08-27 12:34:27.000      2.726405     51.15470 Not classified    white       S.ROP
 2022-08-27 17:34:32.000      2.723473     51.15557     Intertidal cornsilk       S.ROP
 2022-08-28 08:34:55.000      2.757212     51.16779     Intertidal cornsilk       S.ROP

Now I would like to add a legend mapping the habitat types to the specified colour, I tried using scale_colour_manual(), but can't seem to get it right. Any hints?

CodePudding user response：

If you want a legend then map on aesthetics, i.e. move color=... inside aes. And as your color column already contains the colors use scale_color_identity(guide="legend") to add the colors and a legend. Additionally you have to add nice labels when going this way.

For this reason IMHO it would be easier to map habitat directly on the color aes and then use scale_color_manual to set your desired colors.

Moreover, using random_bird$... is strongly discouraged. Just use column names directly when mapping on aes and add the data for the geom_path via data=random_bird.

Note: I added a grey background to the legend keys to make the colors visible.

library(rnaturalearth)
library(ggplot2)

worldmap <- ne_countries(
  scale = "medium", type = "map_units",
  returnclass = "sf"
)
xlim <- range(random_bird$location.long)   0.1 * c(-1, 1)
ylim <- range(random_bird$location.lat)   0.1 * c(-1, 1)

base <- ggplot()  
  geom_sf(data = worldmap)  
  coord_sf(xlim = xlim, ylim = ylim, expand = FALSE)  
  theme_bw()  
  labs(x = "Longitude", y = "Latitude", title = unique(random_bird$colour_ring))  
  theme(plot.title = element_text(hjust = 0.5), legend.key = element_rect(fill = "grey80"))

base  
  geom_path(data = random_bird, aes(x = location.long, y = location.lat, colour = colour))  
  scale_color_identity(guide = "legend")


pal <- random_bird |>
  dplyr::distinct(habitat, colour) |>
  tibble::deframe()

base  
  geom_path(data = random_bird, aes(x = location.long, y = location.lat, colour = habitat))  
  scale_color_manual(values = pal)

DATA

random_bird <- structure(list(timestamp = c(
  " 2022-08-26 14:34:33.000", " 2022-08-26 19:34:30.000",
  "2022-08-27 03:34:31.000", "2022-08-27 05:34:30.000", "2022-08-27 08:34:56.000",
  "2022-08-27 10:34:28.000", "2022-08-27 11:34:28.000", "2022-08-27 12:34:27.000",
  "2022-08-27 17:34:32.000", "2022-08-28 08:34:55.000"
), location.long = c(
  2.732021,
  2.735457, 2.741388, 2.736342, 2.728476, 2.726892, 2.726382, 2.726405,
  2.723473, 2.757212
), location.lat = c(
  51.15316, 51.15966, 51.16101,
  51.15954, 51.15704, 51.15534, 51.15453, 51.1547, 51.15557, 51.16779
), habitat = c(
  "Not classified", "Intertidal", "Intertidal",
  "Intertidal", "Intertidal", "Intertidal", "Not classified", "Not classified",
  "Intertidal", "Intertidal"
), colour = c(
  "white", "cornsilk",
  "cornsilk", "cornsilk", "cornsilk", "cornsilk", "white", "white",
  "cornsilk", "cornsilk"
), colour_ring = c(
  "S.ROP", "S.ROP", "S.ROP",
  "S.ROP", "S.ROP", "S.ROP", "S.ROP", "S.ROP", "S.ROP", "S.ROP"
)), class = "data.frame", row.names = c(NA, -10L))