Add second geom_layer on only one faceted plot with corresponding legend

Question

I'm creating a faceted plot with a first data frame and I add a second geom_point to one of the panel. My problem is I would like to show the corresponding legend for the added point.

To plot the second geom_point on the panel I want, I created a new data frame with the corresponding value and I modified the River column to plot the new geom_point on the correct panel but then the legend is not correct. I would like to have a blue circle in the River section. I learnt a new way to deal with legend on another post, thanks to the people who replied, but here it doesn't work here since this plot is faceted.

df2        <- as_tibble(df1[5,])
df2$River = "Var"

ggplot(data = df1[df1$River != "Roya",], aes(x = Date_plot, y = W_norm, shape = River, col = Type)) +  
  geom_point(size = 3) + 
  scale_shape_manual(values = c(15, 18, 17, 16, 16)) +
  scale_colour_manual(values = c("chocolate1", "darkcyan"), guide = guide_legend(order = 2)) +  
  scale_y_continous("W*") +
  scale_x_date("Years") + 
  geom_point(data = df2, aes(x = Date_plot, y = W_norm, shape = River), colour = "cornflowerblue", size = 3, inherit.aes = F) +
  facet_wrap(vars(River))

Here are the dput of df1 and df2:

structure(list(River = c("Durance", "Durance", "Durance", "Durance", 
"Roya", "Var", "Drac", "Drac", "Drac", "Drac", "Var", "Var", 
"Mareta", "Mareta", "Mareta", "Mareta", "Var"), Type = c("Under restoration", 
"Target", "Under restoration", "Target", "Under restoration", 
"Under restoration", "Under restoration", "Target", "Under restoration", 
"Target", "Target", "Under restoration", "Under restoration", 
"Under restoration", "Target", "Target", "Under restoration"), 
    Date_plot = structure(c(17167, 17167, 15340, 15340, 17532, 
    12784, 14975, 14975, 17532, 17532, 15340, 17532, 12784, 15706, 
    12784, 15706, 15340), class = "Date"), W_norm = c(5.7321, 
    7.9454, 5.1023, 7.0228, 5.0938, 4.7277, 2.7783, 9.303, 7.0742, 
    7.297, 10.2625, 9.5448, 2.83, 5.0009, 3.1914, 3.2644, 4.5448
    )), row.names = c(NA, -17L), class = c("tbl_df", "tbl", "data.frame"
))

structure(list(River = "Var", Type = "Under restoration", Date_plot = structure(17532, class = "Date"), 
    W_norm = 5.0938), row.names = c(NA, -1L), class = c("tbl_df", 
"tbl", "data.frame"))

Why do rivers need to have a separate shape symbol; rivers are distinguished by facet panels and text in the facet strip? — Peter, Apr 12 '21 at 11:28
@Peter Yes it's true, I never thought about it this way! It would be easier then I guess but I would still need to add a blue circle in the legend to show the difference of rivers within the Var plot I think — Jude, Apr 12 '21 at 12:12
There seems to be a `Type` value mismatch between `dput(df2)` `Type = "Under restoration"` and calculated `df2` where `Type = "Witness"`. Could you please explain or correct? — Peter, Apr 12 '21 at 12:29
Sorry it's corrected, I changed the dput for the 1st dataframe dput — Jude, Apr 12 '21 at 12:35
Is there a reason you need to have the `df2` point for the Var river in blue without identifying its `Type` designation: "Under restoration". This seems inconsistent. All other points are either "Under restoration" or "Target". This is not a ggplot issue but may help me understand what the graph is trying to show. According to the logic of the other points the `df2` point for the Var river should be coloured green. — Peter, Apr 12 '21 at 12:40
Yes a little bit of context could help! The point from ```df2``` is the Roya river and I need to plot it on the Var panel because I'm comparing the two rivers, that's why originally I differenciated them with a colour and kept the same shape as the Var river. It's true the ```Type``` is not needed for the Roya river, it just needs to be different than the other points on the Var panel. — Jude, Apr 12 '21 at 12:49
I would suggest the simplest solution is to create a third type: "Roya" which makes your visualisation show your context and avoids the confounding "visual clutter" of different shapes which seem not to add any useful information to the graph. — Peter, Apr 12 '21 at 12:57
Let us [continue this discussion in chat](https://chat.stackoverflow.com/rooms/231034/discussion-between-jude-and-peter). — Jude, Apr 12 '21 at 12:59

score 1 · Accepted Answer · answered Apr 12 '21 at 13:12

Although this does not answer the coding question it may be a solution to the visualisation.

library(ggplot2)

df1$River[5] = "Var"
df1$Type[5] = "Roya, under restoration"

ggplot(data = df1, aes(x = Date_plot, y = W_norm, col = Type)) +  
  geom_point(size = 3) + 
  scale_colour_manual(values = c("chocolate1", "darkcyan", "cornflowerblue")) +  
  labs(y = "W*",
       x = "Years") + 
  facet_wrap(~River)

^{Created on 2021-04-12 by the reprex package (v2.0.0)}

Add second geom_layer on only one faceted plot with corresponding legend

1 Answers1