How to put label on bar plot using ggplot2?

Question

I am beginner in r and I faced to two problems here. I am grateful if anyone can help me.

1. What am I doing wrong that my code does not show me the right distribution in my bar plot? (I am trying to show the survived rate based on sex and class)
2. How can I show the value of each label on top of each bar?

Here is my code:

library(ggplot)

ds <- as.data.frame(Titanic)
color_survived = "#FFA500"
color_dead = "#0000FF"

ds$Sex <- as.factor(ds$Sex)

ds$Survived <- as.factor(ds$Survived)

categorical.ds <- ds  %>%
  select(Sex,
         Class,
         Survived) %>%
  gather(key = "key", value = "value", -Survived)

categorical.ds %>% 
  ggplot(ds ,aes(value)) +
  geom_bar(aes(x        = value, 
               fill     = Survived), 
               alpha    = .2, 
               position = "dodge", 
               color    = "black",
               width    = .7) +
  labs(x = "",
       y = "") +
  theme(
    axis.text.y  = element_blank(),
    axis.ticks.y = element_blank()) +
  facet_wrap(~ key, scales = "free", nrow = 1) +
  scale_fill_manual(values = c(color_survived, color_dead), name   = "Survived", labels = c("Survived", "Dead"))

Here is the plot:

Thank you very much!

Answer 1

here is something, hope it helps!

long story short:

• I think you did something too complex if what you want is just survivors rate, no need to use gather (if you really want to, pivot_longer is more recommended). I used group_by and summarize instead, which is what you need for aggregations.
• Since both graphics haven't the same x, it's simpler to make 2 graphics, one for class, the other for sex. Then combine them with +, having patchwork loaded.
• beware, geom_col is what you want, geom_bar is plotting counts
• Regarding your label issue, just use geom_text

welcome in R commmunity!

##we need the tidyverse package for data manipulation and ggplot
library(tidyverse)
##since we'll make 2 different graphs, patchwork will allow us to combine them
library(patchwork)

#data reproduction
ds <- data.frame(Sex = sample(c("male", "female"), size = 100, replace = TRUE),
                 Survived = sample(c("survived", "dead"), size = 100, replace = TRUE),
                 Class = sample(c("1st", "2nd", "3rd", "crew"), size = 100, replace = TRUE))

#I start with classes : group_by allows us to compute the rate of each group
#summarize uses these group to compute the rate : number of survivors / number in the group (and not NA)
class.ds <-  ds  %>%
  select(Class, Survived) %>%
    group_by(Class) %>%
    summarize(surv_rate = sum(Survived == "survived") / sum(!is.na(Survived)))

#we do the same for the sex
sex.ds <- ds  %>%
  select(Sex, Survived) %>%
  group_by(Sex) %>%
  summarize(surv_rate = sum(Survived == "survived") / sum(!is.na(Survived)))

#now the plots : class is our x axis, the rate is the y axis.
#geom_col is used for the bars, geom_text is for the labels
#of course, you can then add color etc, here I keep it simple
class.plot <- class.ds %>% ggplot(aes(Class, surv_rate)) +
  geom_col() +
  geom_text(aes(label = round(surv_rate, 2)), nudge_y = 0.02)

#same thing for the sex
sex.plot <- sex.ds %>% ggplot(aes(Sex, surv_rate)) +
  geom_col() +
  geom_text(aes(label = round(surv_rate, 2)), nudge_y = 0.02)

#now we just need to group the graphs with patchowork
class.plot + sex.plot