separating geom_point and geom_line in legend

Question

I have a dataset with columns and points superimposed. The plotting is fairly simple but I can't seem to get geom_point to appear as a separate legend. Instead it defaults to combining the two.

I have tried adding the points as separate dataframe, mapping points as factors, and endless functions found on forums. Any hints?

Simplified example below:

library(ggplot2)
Site <- c("A", "B", "C", "A", "B", "C")
Year <- c(2020, 2020, 2020, 2019, 2019, 2019)
Geomean <- c(65,184,27,21,67,51)
Historical.Geomean <- c(76,81,44,76,81,44)

df <- data.frame(Site, Year, Geomean, Historical.Geomean)

df$Year <- as.factor(df$Year)

fw_ecoli_plot <-
  ggplot(df) +
  geom_col(aes(x=Site, y=Geomean, fill=Year), position="dodge") +
  geom_point(aes(x=Site, y=Historical.Geomean),
             color="black", pch=18, show.legend = T) +
  theme_bw()

fw_ecoli_plot

Resulting plot

DaveArmstrong · Answer 1 · 2022-01-24T19:49:38.987

1

How about this:

Site <- c("A", "B", "C", "A", "B", "C")
Year <- c(2020, 2020, 2020, 2019, 2019, 2019)
Geomean <- c(65,184,27,21,67,51)
Historical.Geomean <- c(76,81,44,76,81,44)

df <- data.frame(Site, Year, Geomean, Historical.Geomean)

df$Year <- as.factor(df$Year)

fw_ecoli_plot <-
    ggplot(df) +
    geom_col(aes(x=Site, y=Geomean, fill=Year), position="dodge" ) +
    geom_point(aes(x=Site, y=Historical.Geomean, colour="Historical Mean"),
               pch=18) +
    scale_colour_manual(values="black") + 
    labs(colour="") + 
    theme_bw()

fw_ecoli_plot

edited Jan 24 '22 at 19:49

answered Jan 24 '22 at 19:38

DaveArmstrong

18,377
2
13
25

That's the plot I'm looking for, but I don't see what you changed in the code. I tried the code in your post and produced the same problematic plot. – Oncorhynchus_tshawytscha Jan 24 '22 at 19:42
Sorry, I inadvertently pasted the original code, here's the update. – DaveArmstrong Jan 24 '22 at 19:49
Very tidy. Thanks you much! – Oncorhynchus_tshawytscha Jan 24 '22 at 19:59

TarJae · Accepted Answer · 2022-01-24T19:54:22.550

0

By putting the values in aes, you're basically generating a new factor variable and a legend.

If you put the the values outside of aes -> ggplot won't create a legend. You can try it out:

  df %>% 
    ggplot(aes(x=Site, y=Geomean))+
    geom_col(aes(fill=Year), position="dodge") +
    geom_point(aes(x=Site, y=Historical.Geomean, color="black"), pch=18, show.legend = T) +
    scale_color_manual(values = c('black' = 'black'))+
    guides(fill = guide_legend(override.aes = list(shape = NA)))+
    theme_bw()

edited Jan 24 '22 at 19:54

answered Jan 24 '22 at 19:42

TarJae

72,363
6
19
66

1

Almost there...I would still like to remove the points from the color bar legend. Thanks! – Oncorhynchus_tshawytscha Jan 24 '22 at 19:45
Please see my update! – TarJae Jan 24 '22 at 19:54
1

Thank you! That's it. I was nervous to go down the "override" path, but it looks like that's what it took. Thanks again! – Oncorhynchus_tshawytscha Jan 24 '22 at 19:57

separating geom_point and geom_line in legend

2 Answers2