Tom
Reputation: 2351
Adding p-values to ggplot; ggsignif says it can only handle data with groups that are plotted on the x-axis
I have data as follows, to which I am trying to add p-values:
library(ggplot2)
library(ggsignif)
library(dplyr)
data <- structure(list(treatment = c(0, 1, 0, 1, 0, 1, 0, 1, 0, 1), New_Compare_Truth = c(57,
61, 12, 14, 141, 87, 104, 90, 12, 14), total_Hy = c(135,
168, 9, 15, 103, 83, 238, 251, 9, 15), total = c(285, 305, 60,
70, 705, 435, 520, 450, 60, 70), ratio = c(47.3684210526316,
55.0819672131148, 15, 21.4285714285714, 14.6099290780142, 19.0804597701149,
45.7692307692308, 55.7777777777778, 15, 21.4285714285714), Type = structure(c(2L,
2L, 1L, 1L, 3L, 3L, 5L, 5L, 4L, 4L), .Label = c("A1. Others \nMore \nH",
"A2. Similar \nNorm", "A3. Others \nLess \nH", "B1. Others \nMore \nH",
"B2. Similar \nNorm or \nHigher"), class = "factor"), `Sample Selection` = c("Answers pr",
"Answers pu", "Answers pr", "Answers pu", "Answers pr",
"Answers pu", "Answers pr", "Answers pu", "Answers pr",
"Answers pu"), p_value = c(0.0610371842601616, 0.0610371842601616,
0.346302201593934, 0.346302201593934, 0.0472159407450147, 0.0472159407450147,
0.0018764377521242, 0.0018764377521242, 0.346302201593934, 0.346302201593934
), x = c(2, 2, 1, 1, 3, 3, 5.5, 5.5, 4.5, 4.5)), row.names = c(NA,
-10L), class = c("data.table", "data.frame"))
breaks_labels <- structure(list(Type = structure(c(2L, 1L, 3L, 5L, 4L), .Label = c("A1. Others \nMore \nH",
"A2. Similar \nNorm", "A3. Others \nLess \nH", "B1. Others \nMore \nH",
"B2. Similar \nNorm or \nHigher"), class = "factor"), x = c(2,
1, 3, 5.5, 4.5)), row.names = c(NA, -5L), class = c("data.table",
"data.frame"))
data %>%
ggplot(aes(x = x, y = ratio)) +
geom_col(aes(fill = `Sample Selection`), position = position_dodge(preserve = "single"), na.rm = TRUE) +
geom_text(position = position_dodge(width = .9), # move to center of bars
aes(label=sprintf("%.02f %%", round(ratio, digits = 1)), group = `Sample Selection`),
vjust = -1.5, # nudge above top of bar
size = 4,
na.rm = TRUE) +
# geom_text(position = position_dodge(width = .9), # move to center of bars
# aes(label= paste0("(", ifelse(variable == "Crime = 0", `Observation for Crime = 0`, `Observation for Crime = 1`), ")"), group = `Sample Selection`),
# vjust = -0.6, # nudge above top of bar
# size = 4,
# na.rm = TRUE) +
scale_fill_grey(start = 0.8, end = 0.5) +
scale_y_continuous(expand = expansion(mult = c(0, .1))) +
scale_x_continuous(breaks = breaks_labels$x, labels = breaks_labels$Type) +
theme_bw(base_size = 15) +
xlab("Norm group for corporate Hy") +
ylab("Percentage Compliant Decisions") +
theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
geom_signif(annotation=c("p=0.35", "p=0.06", "p=0.05", "p=0.34", "p=0.00"), y_position = c(30, 40, 55 ,75, 90), xmin=c(0.75,1.75,2.75,3.75,4.75),
xmax=c(1.25,2.25,3.25,4.25,5.25))
For some reason, the last line causes the following error:
Error in f(...) :
Can only handle data with groups that are plotted on the x-axis
Since I am just putting in text and not referring to any variable, I don't really understand why this happens. Can anyone help me out? Without the last line it looks like this:
EDIT: Please note that I would like to keep the space between the third and the fourth column (which is apparently also what caused the problem, see Jared's answer).
Upvotes: 2
Views: 957
Answers (1)
jared_mamrot
Reputation: 26705
Edit
Thanks for clarifying your expected outcome. Here is one way to include geom_signif() annotations without altering the original plot:
library(tidyverse)
library(ggsignif)
data <- structure(list(treatment = c(0, 1, 0, 1, 0, 1, 0, 1, 0, 1), New_Compare_Truth = c(57,
61, 12, 14, 141, 87, 104, 90, 12, 14), total_Hy = c(135,
168, 9, 15, 103, 83, 238, 251, 9, 15), total = c(285, 305, 60,
70, 705, 435, 520, 450, 60, 70), ratio = c(47.3684210526316,
55.0819672131148, 15, 21.4285714285714, 14.6099290780142, 19.0804597701149,
45.7692307692308, 55.7777777777778, 15, 21.4285714285714), Type = structure(c(2L,
2L, 1L, 1L, 3L, 3L, 5L, 5L, 4L, 4L), .Label = c("A1. Others \nMore \nH",
"A2. Similar \nNorm", "A3. Others \nLess \nH", "B1. Others \nMore \nH",
"B2. Similar \nNorm or \nHigher"), class = "factor"), `Sample Selection` = c("Answers pr",
"Answers pu", "Answers pr", "Answers pu", "Answers pr",
"Answers pu", "Answers pr", "Answers pu", "Answers pr",
"Answers pu"), p_value = c(0.0610371842601616, 0.0610371842601616,
0.346302201593934, 0.346302201593934, 0.0472159407450147, 0.0472159407450147,
0.0018764377521242, 0.0018764377521242, 0.346302201593934, 0.346302201593934
), x = c(2, 2, 1, 1, 3, 3, 5.5, 5.5, 4.5, 4.5)), row.names = c(NA,
-10L), class = c("data.table", "data.frame"))
breaks_labels <- structure(list(Type = structure(c(2L, 1L, 3L, 5L, 4L), .Label = c("A1. Others \nMore \nH",
"A2. Similar \nNorm", "A3. Others \nLess \nH", "B1. Others \nMore \nH",
"B2. Similar \nNorm or \nHigher"), class = "factor"), x = c(2,
1, 3, 5.5, 4.5)), row.names = c(NA, -5L), class = c("data.table",
"data.frame"))
annotation_df <- data.frame(signif = c("p=0.35", "p=0.06", "p=0.05", "p=0.34", "p=0.00"),
y_position = c(30, 40, 55 ,75, 90),
xmin = c(0.75,1.75,2.75,4.25,5.25),
xmax = c(1.25,2.25,3.25,4.75,5.75),
group = c(1,2,3,4,5))
data %>%
ggplot(aes(x = x, y = ratio, group = `Sample Selection`)) +
geom_col(aes(fill = `Sample Selection`),
position = position_dodge(preserve = "single"), na.rm = TRUE) +
geom_text(position = position_dodge(width = .9), # move to center of bars
aes(label=sprintf("%.02f %%", round(ratio, digits = 1))),
vjust = -1.5, # nudge above top of bar
size = 4,
na.rm = TRUE) +
scale_fill_grey(start = 0.8, end = 0.5) +
scale_y_continuous(expand = expansion(mult = c(0, .1))) +
scale_x_continuous(breaks = breaks_labels$x, labels = breaks_labels$Type) +
theme_bw(base_size = 15) +
xlab("Norm group for corporate Hy") +
ylab("Percentage Compliant Decisions") +
theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
geom_signif(aes(xmin = xmin,
xmax = xmax,
y_position = y_position,
annotations = signif,
group = group),
data = annotation_df, manual = TRUE)
#> Warning: Ignoring unknown aesthetics: xmin, xmax, y_position, annotations
Created on 2021-07-20 by the reprex package (v2.0.0)
Previous answer
One potential solution to your problem is to plot "Type" on the x axis instead of "x", e.g.
data %>%
ggplot(aes(x = Type, y = ratio)) +
geom_col(aes(fill = `Sample Selection`),
position = position_dodge(preserve = "single"), na.rm = TRUE) +
geom_text(position = position_dodge(width = .9), # move to center of bars
aes(label=sprintf("%.02f %%", round(ratio, digits = 1)),
group = `Sample Selection`),
vjust = -1.5,
size = 4,
na.rm = TRUE) +
scale_fill_grey(start = 0.8, end = 0.5) +
scale_y_continuous(expand = expansion(mult = c(0, .1))) +
theme_bw(base_size = 15) +
xlab("Norm group for corporate Hy") +
ylab("Percentage Compliant Decisions") +
theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
geom_signif(annotation=c("p=0.35", "p=0.06", "p=0.05", "p=0.34", "p=0.00"),
y_position = c(30, 40, 55 ,75, 90),
xmin=c(0.75,1.75,2.75,3.75,4.75),
xmax=c(1.25,2.25,3.25,4.25,5.25))
Upvotes: 2
Related Questions
- Incorrect p-value position on ggplots using rstatix
- Grouped ggplot for adding p-values
- How to resolve problem with ggplot and ggsignif
- Adding ggsignif to plot with numeric x axis results in the error "Can only handle data with groups that are plotted on the x-axis"
- Adding manual p-values to grouped bargraph in ggplot2
- How to add ANOVA one-way p-value to ggplot with ggsignif
- ggplot2: add p-values to the plot
- Plotting p-values with ggpubr in grouped plot results in error
- ggplot and ggsignif error on grouping variable
- Add p-value and r on ggplot [Follow-up]