Keeping unused levels in ggplot2 bar plot does not work properly
R version 4.3.1 library(ggplot2) version 3.4.2
I want to use ggplot2 bar charts to display the percentage frequency of a particular concentration class. I have five concentration classes (=c1 with the options "1_", "2_", "3_", "4_", "5_") for which an event can occur "Early" (=c2 dichotomous with "yes" or "no") or "Late" (=c3 dichotomous with "yes" or "no").
To prevent bars from dropping in the diagram, I have set "scale_x_discrete (drop = FALSE)" in the ggplot.
If an event occurs in one, three, four or all five concentration classes, bar charts are drawn as I imagined them: no bar is drawn for the concentrations for which no event occurred, but all five concentration classes are displayed on the x-axis.
If an event occurs in only two concentration classes, then instead of the two thin bars at the respective concentrations, adjacent bold bars are drawn.
What is the reason for this problem and how can I solve it? Thank you very much in advance for any advice.
This is the code I used:
# build data.frame
c1 <- c("1_", "1_", "2_", "2_", "3_", "4_", "4_", "4_", "5_", "5_")
c2 <- c("yes", "yes", "no", "no", "no", "no", "no", "no", "no", "no")
c3 <- c("no", "no", "yes", "no", "yes", "yes", "no", "no", "no", "yes") # for five bars
c3 <- c("no", "no", "yes", "no", "yes", "no", "no", "no", "no", "yes") # for four bars
c3 <- c("no", "no", "yes", "no", "no", "no", "no", "no", "no", "yes") # for three bars
c3 <- c("no", "no", "no", "no", "no", "no", "no", "no", "no", "yes") # for two bars --> THIS is where the error occurs
c3 <- c("no", "no", "no", "no", "no", "no", "no", "no", "no", "no") # for one bars
dataf <- data.frame(c1, c2, c3)
dataf
# compilation of data + addition of column with event status + reduction of table to relevant columns
bp11 <- dataf %>% mutate(c.Compilation = case_when(c2 == "yes" ~ "EarlyEvent",
c3 == "yes" ~ "LateEvent")) %>%
select(c1, c.Compilation)
bp11
# insert number (n) and percentage share (prop)
bp22 <- bp11 %>% group_by(c1) %>% count(c.Compilation) %>% mutate(prop = n / sum(n) * 100)
# removal of NA
bp33 <- bp22[complete.cases(bp22$c.Compilation), ]
# specify factor level order "Groups"
bp33$c.Compilation = factor(bp33$c.Compilation, levels = c("LateEvent", "EarlyEvent"))
# specify factor level order "Concentration"
bp33$c1 = factor(bp33$c1, levels = c("1_", "2_", "3_", "4_", "5_"))
# calculate n of each concentration group
calcN <- table(bp11$c1)
calcN
# draw stacked barplot
ggplot(bp33,
aes(x = c1, y = prop, fill = c.Compilation)) +
geom_bar(stat = "identity",
position = "stack") +
labs(x = "Concentration (mg/L)",
y = "Event (%)",
fill = "Groups",
title = "Css") +
theme(axis.text = element_text(size = 15, face = "bold"),
axis.title = element_text(size = 15, face = "bold"),
legend.text = element_text(size = 15),
legend.title = element_text(size = 15),
plot.title = element_text(size = 20, hjust = 0.5)) +
theme(legend.position = c(.18, .84)) +
scale_x_discrete(breaks = c("1_", "2_", "3_", "4_", "5_"),
drop = FALSE) +
scale_y_continuous(limit = c(-1, 101), breaks = c(5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100)) +
annotate("text", x = c(0.9, 1.15), y = -1, label = c("n = ", calcN[c("1_")]), size = 5) +
annotate("text", x = c(1.9, 2.15), y = -1, label = c("n = ", calcN[c("2_")]), size = 5) +
annotate("text", x = c(2.9, 3.15), y = -1, label = c("n = ", calcN[c("3_")]), size = 5) +
annotate("text", x = c(3.9, 4.15), y = -1, label = c("n = ", calcN[c("4_")]), size = 5) +
annotate("text", x = c(4.9, 5.15), y = -1, label = c("n = ", calcN[c("5_")]), size = 5) +
scale_fill_grey(start = 0.6, end = 0)
This is the result I get when I run the code:
Solution
This looks like a bug to me (which after some testing occurs only with stat="identity" and when there are no adjacent categories).
NOTE: The issue was already reported and is fixed in the development version of ggplot2. (see below).
A workaround would be to manually complete your dataset using e.g. tidyr::complete:
library(ggplot2)
library(dplyr, warn = FALSE)
library(tidyr)
bp33_complete <- bp33 |>
ungroup() |>
complete(c1, c.Compilation, fill = list(prop = 0))
ggplot(
bp33_complete,
aes(x = c1, y = prop, fill = c.Compilation)
) +
geom_col() +
scale_x_discrete(
breaks = c("1_", "2_", "3_", "4_", "5_"),
drop = FALSE
) +
scale_y_continuous(
breaks = seq(5, 100, 5),
limits = c(-1, 101)
) +
labs(
x = "Concentration (mg/L)",
y = "Event (%)",
fill = "Groups",
title = "Css"
) +
theme(
axis.text = element_text(size = 15, face = "bold"),
axis.title = element_text(size = 15, face = "bold"),
legend.text = element_text(size = 15),
legend.title = element_text(size = 15),
plot.title = element_text(size = 20, hjust = 0.5)
) +
theme(legend.position = c(.18, .84)) +
geom_text(
aes(
label = after_stat(paste0("n = ", count)),
y = after_stat(-1),
fill = NULL
),
data = bp11,
stat = "count",
size = 5
) +
scale_fill_grey(start = 0.6, end = 0)
DATA
c1 <- c("1_", "1_", "2_", "2_", "3_", "4_", "4_", "4_", "5_", "5_")
c2 <- c("yes", "yes", "no", "no", "no", "no", "no", "no", "no", "no")
# for two bars --> THIS is where the error occurs
c3 <- c("no", "no", "no", "no", "no", "no", "no", "no", "no", "yes")
dataf <- data.frame(c1, c2, c3)
EDIT Here is a refactored code which uses only one dataset, keeps the NAs (but drops them for the geom_col) and uses geom_text with stat="summary" to compute the n= which should also work for cases with n=0.
library(ggplot2)
library(dplyr, warn = FALSE)
library(tidyr)
# Drop one c1 category
dataf <- filter(dataf, c1 != "4_")
bp33 <- dataf %>%
mutate(c.Compilation = case_when(
c2 == "yes" ~ "EarlyEvent",
c3 == "yes" ~ "LateEvent"
)) %>%
count(c1, c.Compilation) %>%
mutate(prop = n / sum(n) * 100, .by = c1) |>
mutate(
c.Compilation = factor(c.Compilation,
levels = c("LateEvent", "EarlyEvent")
),
c1 = factor(c1,
levels = c("1_", "2_", "3_", "4_", "5_")
)
) |>
complete(c1, c.Compilation, fill = list(prop = 0, n = 0))
ggplot(
bp33,
aes(x = c1, y = prop)
) +
geom_col(
data = ~ filter(.x, !is.na(c.Compilation)),
aes(fill = c.Compilation)
) +
scale_x_discrete(
breaks = c("1_", "2_", "3_", "4_", "5_"),
drop = FALSE
) +
scale_y_continuous(
breaks = seq(5, 100, 5),
limits = c(-1, 101)
) +
labs(
x = "Concentration (mg/L)",
y = "Event (%)",
fill = "Groups",
title = "Css"
) +
theme(
axis.text = element_text(size = 15, face = "bold"),
axis.title = element_text(size = 15, face = "bold"),
legend.text = element_text(size = 15),
legend.title = element_text(size = 15),
plot.title = element_text(size = 20, hjust = 0.5)
) +
theme(legend.position = c(.18, .84)) +
geom_text(
aes(
label = after_stat(paste0("n = ", y)),
y = stage(n, after_stat = -1),
fill = NULL,
group = c1
),
stat = "summary",
fun = sum,
size = 5
) +
scale_fill_grey(
start = 0.6, end = 0
)
Update The issue was already reported here and here and seems to be fixed in the development version of ggplot2:
Using ggplot2 3.4.4:
library(ggplot2)
packageVersion("ggplot2")
#> [1] '3.4.4'
dat <- data.frame(
x = factor(c("A", "E"), levels = LETTERS[1:5])
)
ggplot(dat, aes(x, y = 1)) +
geom_bar(stat = "identity") + # or geom_col
scale_x_discrete(drop = FALSE)
Created on 2023-12-30 with reprex v2.0.2
Using the development version:
library(ggplot2)
packageVersion("ggplot2")
#> [1] '3.4.4.9000'
dat <- data.frame(
x = factor(c("A", "E"), levels = LETTERS[1:5])
)
ggplot(dat, aes(x, y = 1)) +
geom_bar(stat = "identity") + # or geom_col
scale_x_discrete(drop = FALSE)
- grid.arrange's relative height argument didn't works well in RMarkdown
- How apply a function to every element of a matrix?
- Polar plot in R using ggplot2
- Find total length of time in multiple values of a dataset that overlaps with a given period of time
- Faster than R's `simplify2array`?
- What is the most simple way to label a data.table frequency chart in R?
- How to show dates in a secondary x-axis while the primary x-axis is numerical?
- online compiler for r-markdown document
- Unecessarily large column widths in HTML tables published by R Quarto when the table header contains special characters
- Bootstrapping a vector of results, by group in R
- Speeding Up Monte Carlo Simulations for Weather-Dependent Claims
- Is there an R class which allows to list 2 elements but shows only the first of them by default when called?
- Convert day of week number to weekday name in R
- How do you use "<<-" (scoping assignment) in R?
- Using reactivePoll to monitor and process file changes
- How to find out which graph is most similar to another graph?
- ggplot2: rectangular background behind axis text in circular bar plot
- How to simulate button click using rvest
- CMIP6, automating by month
- Covert dummy variables to single categorical in R?
- Animated reducing bar chart with gganimate
- Create an object class matrix/array from data frame
- How to add a vector of names to `sort_by()` in base R
- Among many subfolders, find CSVs starting with a given string and within ZIPs or not, and merge them while adding their names in a new column with R
- How to add a greek letter to an axis label to a 3D plot in plotly?
- bigrquery + dplyr error when using functions and sourcing from separate R script
- Shiny App Error: /v1/applications/ 400 - Validation Error Execution halted
- Rupee Unicode replacement not working in R
- Mutate multiple variable to create multiple new variables
- How to use the Stata equivalent of a foreach loop in R with mutate()