Display all levels in legend of ggplotly object

Question

How do I display all legends in R ggplotly? As of now it only displays the legends for matching data.

Here is what I have done so far:

library(ggplot2)
library(plotly)

schoolname <- c("Xavier", "Dakota")
value <- c(34,50)
df <- data.frame(schoolname,value)

ggplot(df, aes(x=schoolname, y=value, 
               fill = cut(value, breaks =c(0,34,69,100), 
                          labels = c("Failed -     <65%", 
                                     "Pass - 65%", 
                                     "Excellent - 80%")))) +
  geom_bar(stat = "identity",position="dodge", width=0.5, color = "#333333") +
  geom_text(aes(x=schoolname, y=value + 4.1, 
                label = paste0(value), tooltip = NULL), 
            inherit.aes = F, color='black', 
            position = position_dodge2(2), size=4, vjust=1.5) + 
  xlab("School Name") +
  ylab("Result (%)") +
  theme(plot.title = element_text(size = 18, hjust=0, vjust=0)) +
  scale_y_continuous(limits = c(0,100))+
  scale_fill_manual(values = c("Failed - <65%" = "#ea9999", 
                               "Pass - 65%" = "#ffc8aa", 
                               "Excellent- 80%" = "#d4edbc")) + 
  theme(legend.position = "top") +
  guides(fill = guide_legend(title = "Result", override.aes = list(size = 4))) -> p

M-- · Accepted Answer · 2024-12-03 17:02:01Z

1

We can add a dummy row (schoolname == NA) with "Excellent" rating, then that level is shown in the legend without getting plotted. Moreover, you had a typo where your levels in cut() did not match the ones in scale_fill_...(). Also note that your breaks in cut() do not match what you are showing in your legend.

library(dplyr)
library(ggplot2)
library(plotly)

df %>% 
  add_row(value = 100) %>% ## adding the dummy row
ggplot(aes(x=schoolname, y=value, fill = cut(value, breaks =c(0,34,69,100), 
                                             labels = c("Failed - <65%", 
                                                        "Pass - 65%", 
                                                        "Excellent - 80%")))) +
  geom_bar(stat = "identity", position="dodge", width=0.5, color = "#333333") +
  geom_text(aes(x=schoolname, y=value + 4.1, 
                label = paste0(value), tooltip = NULL), 
            inherit.aes = F, color='black', 
            position = position_dodge2(2), size=4, vjust=1.5) + 
  xlab("School Name") +
  ylab("Result (%)") +
  theme(plot.title = element_text(size = 18, hjust=0, vjust=0)) +
  scale_y_continuous(limits = c(0,100)) +
  scale_x_discrete(limits = df$schoolname) +
  scale_fill_manual(values = c("Failed - <65%" = "#ea9999", 
                               "Pass - 65%" = "#ffc8aa", 
                               "Excellent - 80%" = "#d4edbc"), 
                    drop = FALSE) + 
  theme(legend.position = "top") +
  guides(fill = guide_legend(title = "Result", override.aes = list(size = 4))) -> p

ggplotly(p)

^{Created on 2024-07-14 with reprex v2.0.2}

Initially I thought of closing this as a duplicate of Show entire legend even when a category is not displayed in the graph, but this needs a different approach since the data is not presently available for every level.

edited Dec 3, 2024 at 17:02

answered Jul 14, 2024 at 5:33

M--

33.6k12 gold badges74 silver badges115 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

stefan Over a year ago

Besides the typos the main issue is related to the conversion to a plotly object and hence is more likely a duplicate of Prevent ggplotly to drop unused factor levels from legend which also adds a dummy row. (:

M-- Over a year ago

@stefan I didn't find that dupe-target, and it is a bit different from here, but close enough to close this I guess. If you do so, I won't contest :)

M-- · Accepted Answer · 2024-07-14 05:26:24Z

0

If I make the labels in your cut() function exactly match the labels in scale_fill_manual() and I add drop = FALSE to scale_fill_manual(), I do see all the labels in the legend. Make sure the number of spaces in each label is exactly the same between the two sets. For example, check "Failed - <65%" in your code. There is also some spurious text in your scale_fill_manual(). It starts with "[enter image description here]".

The unpopulated level is shown with a light gray color in the legend. I suppose that is intentional on the part of the ggplot2 designers.

p <- ggplot(df, aes(x=schoolname, y=value, 
                    fill = cut(value, breaks =c(0,34,69,100), 
                               labels = c("Failed - <65%", "Pass - 65%", "Excellent - 80%"))))+
  geom_bar(stat = "identity",position="dodge", width=0.5, color = "#333333") +
  geom_text(aes(x=schoolname, y=value + 4.1, label = paste0(value), tooltip = NULL), inherit.aes = F,   color='black', position=position_dodge2(2), size=4, vjust=1.5)+ 
  xlab("School Name") +
  ylab("Result (%)") +
  theme(plot.title = element_text(size = 18, hjust=0, vjust=0)) +
  scale_y_continuous(limits = c(0,100))+
  scale_fill_manual(values = c("Failed - <65%" = "#ea9999", 
                               "Pass - 65%" = "#ffc8aa", 
                               "Excellent - 80%" = "#d4edbc"), drop = FALSE) + 
  theme(legend.position = "top")+
  guides(fill = guide_legend(title = "Result", override.aes = list(size = 4)))

print(p)

edited Jul 14, 2024 at 5:26

M--

33.6k12 gold badges74 silver badges115 bronze badges

answered Jul 14, 2024 at 2:09

FJCC

1,1212 silver badges3 bronze badges

2 Comments

M-- Over a year ago

It's shown as grey because it's NA! Yes, that's intentional, but not what OP is seeking.

stefan Over a year ago

Besides drop=FALSE you also have to add show.legend=TRUE in ggplot2 >= 3.5.0 to show a legend key for unused levels. See GGplot Legend Showing Blank Fill for Missing Factor Level. However, while this works for ggplot2 the key still gets dropped when converting to a plotly object.

Collectives™ on Stack Overflow

Display all levels in legend of ggplotly object

2 Answers 2

2 Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related