How to format legend for dual axis plot in ggplot2

October 13, 2022

I have the following code in R to make a slightly complex dual axis plot using ggplot. My plot is almost complete, however I’m having an issue with my legend. I’ve noticed that ggplot creates two seperate legends if one combines two different geoms into one chart – is there a way in which I can merge the two legends into one? Also, my second legend has two entries in it and I only need the first one (Overall Balance (% of GDP, RHS)), is there a way to remove the second Overall Balance (N$ Million)?

dat <- structure(list(Category = c("Overall Balance (N$ Million)", "Overall Balance (% of GDP, RHS)", 
"Deficit Target (% of GDP, RHS)", "Overall Balance (N$ Million)", 
"Overall Balance (% of GDP, RHS)", "Deficit Target (% of GDP, RHS)", 
"Overall Balance (N$ Million)", "Overall Balance (% of GDP, RHS)", 
"Deficit Target (% of GDP, RHS)", "Overall Balance (N$ Million)", 
"Overall Balance (% of GDP, RHS)", "Deficit Target (% of GDP, RHS)", 
"Overall Balance (N$ Million)", "Overall Balance (% of GDP, RHS)", 
"Deficit Target (% of GDP, RHS)", "Overall Balance (N$ Million)", 
"Overall Balance (% of GDP, RHS)", "Deficit Target (% of GDP, RHS)", 
"Overall Balance (N$ Million)", "Overall Balance (% of GDP, RHS)", 
"Deficit Target (% of GDP, RHS)"), Year = c("2018/19", "2018/19", 
"2018/19", "2019/20", "2019/20", "2019/20", "2020/21", "2020/21", 
"2020/21", "2021/22 (est)", "2021/22 (est)", "2021/22 (est)", 
"2022/23 (proj)", "2022/23 (proj)", "2022/23 (proj)", "2023/24 (proj)", 
"2023/24 (proj)", "2023/24 (proj)", "2024/25 (proj)", "2024/25 (proj)", 
"2024/25 (proj)"), Value = c(-9226, -5.09893041491584, -3, -8918, 
-4.97862890962981, -3, -14197, -8.04324429255841, -3, -16138, 
-8.8, -3, -12658, -6.5, -3, -10418, -5, -3, -10281, -4.7, -3)), class = c("tbl_df", 
"tbl", "data.frame"), row.names = c(NA, -21L))

library(tidyverse)

ggplot() +
  geom_col(data = subset(dat, Category %in% c("Overall Balance (N$ Million)")),
           mapping = aes(x = Year, y = Value, fill = Category, col = Category),
           position = position_dodge()) +
  geom_line(data = subset(dat, Category == "Overall Balance (% of GDP, RHS)"),
            mapping = aes(x = Year, y = (Value*1500), group = Category,color = Category), size = 2) +
  scale_fill_manual(values = c('#7c3042','#6d1d26')) +
  scale_color_manual(values = c('#c7af76','#c7af76')) +
  theme_light() +
  labs(x = '') +
  scale_y_continuous(breaks = seq(-22000,3000,by = 5000),
                     name = 'Overall Balance (N$ Million)',sec.axis= sec_axis(~./1500, name="Overall Balance (% of GDP)", breaks = seq(-10,0, by = 2))) +
  theme(legend.title = element_blank())

>Solution :

This issue is independent of the secondary y-axis. The issue is that in geom_col() you map the colour aesthetic, whereas this layer should not get an entry in the colour legend (just the fill legend only). You can fix it by defining the colour outside the aes() mapping of the layer.

dat <- structure(list(Category = c("Overall Balance (N$ Million)", "Overall Balance (% of GDP, RHS)", 
                                   "Deficit Target (% of GDP, RHS)", "Overall Balance (N$ Million)", 
                                   "Overall Balance (% of GDP, RHS)", "Deficit Target (% of GDP, RHS)", 
                                   "Overall Balance (N$ Million)", "Overall Balance (% of GDP, RHS)", 
                                   "Deficit Target (% of GDP, RHS)", "Overall Balance (N$ Million)", 
                                   "Overall Balance (% of GDP, RHS)", "Deficit Target (% of GDP, RHS)", 
                                   "Overall Balance (N$ Million)", "Overall Balance (% of GDP, RHS)", 
                                   "Deficit Target (% of GDP, RHS)", "Overall Balance (N$ Million)", 
                                   "Overall Balance (% of GDP, RHS)", "Deficit Target (% of GDP, RHS)", 
                                   "Overall Balance (N$ Million)", "Overall Balance (% of GDP, RHS)", 
                                   "Deficit Target (% of GDP, RHS)"), Year = c("2018/19", "2018/19", 
                                                                               "2018/19", "2019/20", "2019/20", "2019/20", "2020/21", "2020/21", 
                                                                               "2020/21", "2021/22 (est)", "2021/22 (est)", "2021/22 (est)", 
                                                                               "2022/23 (proj)", "2022/23 (proj)", "2022/23 (proj)", "2023/24 (proj)", 
                                                                               "2023/24 (proj)", "2023/24 (proj)", "2024/25 (proj)", "2024/25 (proj)", 
                                                                               "2024/25 (proj)"), Value = c(-9226, -5.09893041491584, -3, -8918, 
                                                                                                            -4.97862890962981, -3, -14197, -8.04324429255841, -3, -16138, 
                                                                                                            -8.8, -3, -12658, -6.5, -3, -10418, -5, -3, -10281, -4.7, -3)), class = c("tbl_df", 
                                                                                                                                                                                      "tbl", "data.frame"), row.names = c(NA, -21L))

library(tidyverse)

ggplot() +
  geom_col(data = subset(dat, Category %in% c("Overall Balance (N$ Million)")),
           mapping = aes(x = Year, y = Value, fill = Category),
           colour = "#c7af76",
           position = position_dodge()) +
  geom_line(data = subset(dat, Category == "Overall Balance (% of GDP, RHS)"),
            mapping = aes(x = Year, y = (Value*1500), group = Category,color = Category), size = 2) +
  scale_fill_manual(values = c('#7c3042','#6d1d26')) +
  scale_color_manual(values = c('#c7af76','#c7af76')) +
  theme_light() +
  labs(x = '') +
  scale_y_continuous(breaks = seq(-22000,3000,by = 5000),
                     name = 'Overall Balance (N$ Million)',sec.axis= sec_axis(~./1500, name="Overall Balance (% of GDP)", breaks = seq(-10,0, by = 2))) +
  theme(legend.title = element_blank())