Legend is not shown using ggplot2 despite using color within aes()

Question

I have a problem making the plot legend to show. I've checked other posts and most of them mention that color should be defined within the aes() of the ggplot.

So I have the following dataset:

  date     `N Rx AFTN` `N Tx AFTN`  SumAFTN `N Rx AMHS` `N Tx AMHS`  SumAMHS   SUM_Tx SUM_Total   SUM_Rx
  <chr>          <int>       <int>    <dbl>       <int>       <int>    <dbl>    <dbl>     <dbl>    <dbl>
1 2018_M12    22844731    36682853 59527584     5932852     5942479 11875331 42625332  71402915 28777583
2 2019_M12    31858229    52049649 83907878    23412883    25242891 48655774 77292540 132563652 55271112
3 2020_M12    12788348    27191283 39979631    14769682    18274489 33044171 45465772  73023802 27558030
4 2021_M12    11718972    36313167 48032139    22200241    24076705 46276946 60389872  94309085 33919213
5 2022_M12    13362373    46255935 59618308    32268850    31050963 63319813 77306898 122938121 45631223
6 2023_M12    14479914    50957248 65437162    35808014    34036231 69844245 84993479 135281407 50287928

and I'm plotting using:

    print(ggplot(all_df))+
      geom_point(aes(x=date,y=SUM_Total/1000000))+
      geom_line(aes(x=date,y=SUM_Total/1000000,color=SUM_Total),color="blue",group=1)+
      geom_line(aes(x=date,y=SUM_Rx/1000000, color=SUM_Rx),color="green",group=1) +
      geom_line(aes(x=date,y=SUM_Tx/1000000,color=SUM_Tx),color="red",group=1)

The resulting plot is

where no legend is shown.

What am I missing?

The output of dput(all_df) is :

structure(list(date = c("2018_M12", "2019_M12", "2020_M12", "2021_M12", "2022_M12", "2023_M12"), N Rx AFTN = c(22844731L, 31858229L, 12788348L, 11718972L, 13362373L, 14479914L), N Tx AFTN = c(36682853L, 52049649L, 27191283L, 36313167L, 46255935L, 50957248L), SumAFTN = c(59527584, 83907878, 39979631, 48032139, 59618308, 65437162), N Rx AMHS = c(5932852L, 23412883L, 14769682L, 22200241L, 32268850L, 35808014L), N Tx AMHS = c(5942479L, 25242891L, 18274489L, 24076705L, 31050963L, 34036231L), SumAMHS = c(11875331, 48655774, 33044171, 46276946, 63319813, 69844245), SUM_Rx = c(28777583, 55271112, 27558030, 33919213, 45631223, 50287928), SUM_Tx = c(42625332, 77292540, 45465772, 60389872, 77306898, 84993479), SUM_Total = c(71402915, 132563652, 73023802, 94309085, 122938121, 135281407)), row.names = c(NA, -6L), class = c("tbl_df", "tbl", "data.frame"), .internal.selfref = <pointer: 0x5613f6742840>, sorted = "date")

Answer 1

When you assign colors using color argument directly within the geom_line() function, ggplot interprets these as instructions to use a specific color rather than to map data to an aesthetic, which is what generates legends. In order to generate a legend based on the data, you need to move the color argument into the aes() function. Here I use scale_color_manual().

ggplot(all_df) +
  geom_point(aes(x = date, y = SUM_Total / 1000000)) +
  geom_line(aes(x = date, y = SUM_Total / 1000000, color = "Total"), group = 1) +
  geom_line(aes(x = date, y = SUM_Rx / 1000000, color = "Rx"), group = 1) +
  geom_line(aes(x = date, y = SUM_Tx / 1000000, color = "Tx"), group = 1) +
  scale_color_manual(values = c("Total" = "blue", "Rx" = "green", "Tx" = "red"),
                     name = "Message Types", labels = c("Total", "Received", "Transmitted")) +
  labs(title = "Plot Title", x = "Date/Time", y = "Messages (in millions)") +
  theme(axis.text = element_text(size = 8), 
        axis.title = element_text(size = 8),
        plot.title = element_text(size = 10),
        axis.text.x = element_text(angle = 45))

Answer 2

And here's a tidy solution. I leave you to sort out the finer details of the formatting for yourself.

all_df %>% 
  pivot_longer(starts_with("SUM_")) %>% 
  mutate(plotDate = as.numeric(str_sub(date, 1, 4))) %>% 
  ggplot() +
  geom_line(aes(x = plotDate,y = value/1000000, color = name))

pivot_longer both filters the data frame to include only the "columns" you need and produces a data frame with one row per value of date per column in the original data frame and three columns: date, name and value. This makes it tidy.

A numeric value is needed on the x axis to make geom_line and aes(colour=) interact correctly: hence the call to mutate.

Answer 3

Thank you all for your support.

It seems to me that, end of the day, there was an issue with the RStudio I was using (.deb for Ubuntu linux).

Ofcourse there were mistakes in the code I posted above regarding the color outsise the aesthetics -I had tried that as well-, but after restarting the RStudio the Legend appeared.

Thanks again