How to make stacked bar chart using ggplot2? [duplicate] - r

Question

This question already has answers here:
Create stacked barplot where each stack is scaled to sum to 100%
(5 answers)
Closed 10 months ago.
So I am having trouble making a stacked bar chart showing proportion of cases vs deaths.
This is the data:
df <- structure(list(Date = structure(c(19108, 19108, 19108, 19108,
19108, 19108, 19108, 19108, 19108, 19108), class = "Date"), Country = c("US",
"India", "Brazil", "France", "Germany", "United Kingdom", "Russia",
"Korea, South", "Italy", "Turkey"), Confirmed = c(81100599L,
43065496L, 30378061L, 28605614L, 24337394L, 22168390L, 17887152L,
17086626L, 16191323L, 15023662L), Recovered = c(0L, 0L, 0L, 0L,
0L, 0L, 0L, 0L, 0L, 0L), Deaths = c(991940L, 523654L, 663108L,
146464L, 134489L, 174778L, 367692L, 22466L, 162927L, 98720L),
Active = c(80108659L, 42541842L, 29714953L, 28459150L, 24202905L,
21993612L, 17519460L, 17064160L, 16028396L, 14924942L)), row.names = c(163539L,
163431L, 163375L, 163414L, 163418L, 163537L, 163496L, 163444L,
163437L, 163533L), class = "data.frame")
and I want to generate something that looks like this except with proportions of deaths vs cases.

Answer 1

This is a modification of #Allan Cameron's answer with adding the percent label and some other different approaches:
library(tidyverse)
library(scales)
df %>%
rename_with(., ~str_replace_all(., 'top10.', '')) %>%
pivot_longer(
cols = -Country,
names_to = "Status",
values_to = "value",
values_transform = list(value = as.integer)
) %>%
mutate(Status = fct_rev(fct_infreq(Status))) %>%
group_by(Country) %>%
mutate(pct= prop.table(value) * 100) %>%
ggplot(aes(x= Country, y = pct, fill=Status)) +
geom_col(position = position_fill())+
scale_fill_manual(values = c("#ff34b3", "#4976ff")) +
scale_y_continuous(labels = scales::percent)+
ylab("Percentage") +
geom_text(aes(label=paste0(sprintf("%1.1f", pct),"%")),
position=position_fill(vjust = 0.1)) +
ggtitle("Your Title")

Answer 2

I had to use OCR to convert the image of your data into actual data I could use. It's far better to include your data as text for this reason.
The plot is not particularly informative because the percentages are low, and difficult to read, but in any case, you can do it like this:
library(tidyverse)
p <- df %>%
mutate(top10.Confirmed = top10.Confirmed - top10.Deaths,
top10.Country = factor(top10.Country, top10.Country)) %>%
rename(Country = top10.Country,
Survived = top10.Confirmed,
Died = top10.Deaths) %>%
pivot_longer(-Country, names_to = "Outcome", values_to = "Count") %>%
mutate(Outcome = factor(Outcome, c("Survived", "Died"))) %>%
ggplot(aes(Country, Count, fill = Outcome)) +
geom_col(position = "fill") +
scale_fill_manual(values = c("#4976ff", "#ff34b3")) +
scale_y_continuous(labels = scales::percent) +
labs(title = "Covid outcomes by country", y = "Percent")
p
To make it easier to read, you could zoom into the bottom:
p + coord_cartesian(ylim = c(0, 0.05))
Data in reproducible format
df <- structure(list(top10.Country = c("US", "India", "Brazil", "France",
"Germany", "United Kingdom", "Russia", "Korea, South", "Italy",
"Turkey"), top10.Confirmed = c(81100599L, 43065496L, 30378061L,
28605614L, 24337394L, 22168390L, 17887152L, 17086626L, 16191323L,
15023662L), top10.Deaths = c(991940L, 523654L, 663108L, 146464L,
134489L, 174778L, 367692L, 22466L, 162927L, 98720L)), class = "data.frame",
row.names = c("1", "2", "3", "4", "5", "6", "7", "8", "9", "10"))
df
#> top10.Country top10.Confirmed top10.Deaths
#> 1 US 81100599 991940
#> 2 India 43065496 523654
#> 3 Brazil 30378061 663108
#> 4 France 28605614 146464
#> 5 Germany 24337394 134489
#> 6 United Kingdom 22168390 174778
#> 7 Russia 17887152 367692
#> 8 Korea, South 17086626 22466
#> 9 Italy 16191323 162927
#> 10 Turkey 15023662 98720
Created on 2022-05-01 by the reprex package (v2.0.1)