I have a dataframe that looks something like this:
Columns with data:
Sex Var1 Var2 Var3
Male Yes No Yes
Female No Yes No
Male No No Yes
Female Yes Yes No
I want to create a barplot with two bars for each Variable (one for Males, one for Females) in the X axis and the "Yes" count in the Y axis and I don't really know how. Any help would be much appreciated.
Here is one with base R. Get the count of 'Yes' with rowSums on a logical matrix selecting only the 'Var' columns, then do a group by 'Sex' to summarise the count by Sex with rowsum) and use barplot
barplot(t(rowsum(rowSums(df1[-1] == 'Yes'), df1$Sex)))
Or if we need a group by barplot, change it to
barplot(t(rowsum(+(df1[-1] == 'Yes'), df1$Sex)), beside = TRUE,
legend = TRUE, col = c('red', 'blue', 'green'))
Or if we prefer ggplot, reshape to 'long' format with pivot_longer (from tidyr), get a group_by, summarise to return the count of 'Yes' and use ggplot
library(dplyr)
library(tidyr)
library(ggplot2)
df1 %>%
pivot_longer(cols = -Sex) %>%
group_by(Sex) %>%
summarise(n = sum(value == 'Yes')) %>%
ggplot(aes(x = Sex, y = n)) +
geom_col()
For a bar for each 'Var'
df1 %>%
pivot_longer(cols = -Sex) %>%
group_by(Sex, name) %>%
summarise(n = sum(value == 'Yes'), .groups = 'drop') %>%
ggplot(aes(x = Sex, y = n, fill = name)) +
geom_col(position = 'dodge')
-output
df1 <- structure(list(Sex = c("Male", "Female", "Male", "Female"),
Var1 = c("Yes",
"No", "No", "Yes"), Var2 = c("No", "Yes", "No", "Yes"), Var3 = c("Yes",
"No", "Yes", "No")), class = "data.frame", row.names = c(NA,
-4L))
this should get you going.
data for reproducible example
df <- tribble(
~Sex, ~Var1, ~Var2, ~Var3
,"Male", "Yes", "No", "Yes"
,"Female", "No", "Yes", "No"
,"Male", "No", "No", "Yes"
,"Female", "Yes", "Yes", "No"
)
summarise the data set
library(dplyr)
df <- df %>%
group_by(Sex) %>%
summarise(Var1_Y = sum(Var1 == "Yes")
,Var2_Y = sum(Var2 == "Yes")
,Var3_Y = sum(Var3 == "Yes")
)
this gives you
df
# A tibble: 2 x 4
Sex Var1_Y Var2_Y Var3_Y
<chr> <int> <int> <int>
1 Female 1 2 0
2 Male 1 0 2
ggplot works best with "long" data. for this we use pivot_longer().
library(tidyr)
df <- df %>%
pivot_longer(cols = -Sex, names_to = "Var", values_to = "Val")
making it a long data frame
df
# A tibble: 6 x 3
Sex Var Val
<chr> <chr> <int>
1 Female Var1_Y 1
2 Female Var2_Y 2
3 Female Var3_Y 0
4 Male Var1_Y 1
5 Male Var2_Y 0
6 Male Var3_Y 2
with this you can easily plot with ggplot For totals use geom_col() and not geom_bar()
library(ggplot)
df %>%
ggplot(aes(x = Var, y = Val, fill = Sex)) +
geom_col(position = "dodge") # dodge puts the bars next to each other
Here is another barplot presentation using data.table + ggplot2, i.e.,
melt(setDT(df), id.var = "Sex")[
,
.(N = sum(value == "Yes")),
.(Sex, variable)
] %>%
ggplot(aes(x = Sex, y = N, fill = variable)) +
geom_bar(stat = "identity")
which gives
