使用 ggplot2 在条形图上添加症状表？

Question

I want to add my symptom names on each bar corresponding to each colour, in ggplot2, R. I want the names on each bar to be aligned with the colour and I pressume a parameter with regards to the size of each symptom word has to be passed, yet do not know how.

I have tried but did not succeed. This is how it looks with what I have tried:

a fake data can be found to this link:

https://github.com/gabrielburcea/stackoverflow_fake_data/blob/master/labels_symptoms_ontop_of_bar_data.csv

And this is the code I have tried:

plot_adjusted_rates <- ggplot2::ggplot(fake_data, 
                                       ggplot2::aes(country, value)) +
  ggplot2::coord_flip() +
  ggplot2::geom_bar(ggplot2::aes(fill = symptoms), width = 0.4,
                    position = position_dodge(width = 0.5), stat = "identity") +
  geom_text(aes(label = symptoms), nudge_x = c(0.22, -0.22)) + 
  jcolors::scale_color_jcolors(palette = "pal12" )+ 
  ggplot2::labs(title = title,
                subtitle = "\nNote: Adjusted rates for symptoms in responders tested  in %",
                x = "Symptoms in /Countries", y = "Percentage", caption = "Source: Your.md Data") +
  ggplot2::theme(axis.title.y = ggplot2::element_text(margin = ggplot2::margin(t = 0, r = 21, b = 0, l = 0)),
                 plot.title = ggplot2::element_text(size = 12, face = "bold"),
                 plot.subtitle = ggplot2::element_text(size = 10),
                 legend.position = "bottom" , legend.box = "horizontal") +
  ggplot2::theme_bw()


plot_adjusted_rates

Answer 1

Here is a general code chunk that would solve your problem:

# load librariy
library(dplyr)
library(ggplot2)

# load data
data_url = 'https://raw.githubusercontent.com/gabrielburcea/stackoverflow_fake_data/master/labels_symptoms_ontop_of_bar_data.csv'
fake_data = read.csv(data_url)

# plot
plot = ggplot(fake_data, aes(x = country, y = value, fill = symptoms)) +
  geom_bar(stat = "identity", show.legend = FALSE,
           width = 0.4, position = position_dodge(width = 0.5)) +
  coord_flip() + 
  geom_text(aes(label = symptoms), size = 3,
            hjust = -0.05, position = position_dodge2(width = 0.5))

Output:

However, as you can see, the labels from geom_text() are overlapping. It happens because the same country and symptoms have more than one value in your fake data. The country A shows a particular "problem", because the same symptom have pretty different values (24.05 and 44.05). It makes me think if you shouldn't clean your data... Take a look:

dplyr::filter(fake_data, symptoms == 'loss_appetite')

Output:

   country      symptoms    value
1        A loss_appetite 24.05464
2        A loss_appetite 24.05464
3        A loss_appetite 24.05464  <- 24.05
4        A loss_appetite 44.05464  <- 44.05
5        B loss_appetite 31.25430
6        B loss_appetite 31.25430
7        B loss_appetite 31.25430
8        B loss_appetite 31.25430
9        C loss_appetite 32.44539
10       C loss_appetite 32.44539
11       C loss_appetite 32.44539
12       C loss_appetite 32.44539
13       D loss_appetite 36.52090
14       D loss_appetite 36.52090
15       D loss_appetite 36.52090
16       D loss_appetite 36.52090
17       E loss_appetite 20.65789

To avoid the overlapping text issue, as well as the data "mistake" pointed out, you can use groupy() and summarise() , in order to sort out only the maximum values for each group and place the text on the top of each bar.

Thus, this code should plot the data as you wish:

# create label data
fake_text = fake_data %>%
  group_by(country, symptoms) %>%
  summarize(max = max(value))

# plot
ggplot(fake_data, aes(x = country, y = value, fill = symptoms)) +
  geom_bar(stat = "identity", show.legend = FALSE,
           width = 0.4, position = position_dodge(width = 0.5)) +
  coord_flip() + 
  geom_text(aes(label = symptoms), size = 3,
            hjust = -0.05, position = position_dodge2(width = 0.5))

Output: