Let's say I have the following data frame:
library(ggplot2)
set.seed(101)
n=10
df<- data.frame(delta=rep(rep(c(0.1,0.2,0.3),each=3),n), metric=rep(rep(c('P','R','C'),3),n),value=rnorm(9*n, 0.0, 1.0))
My goal is to do a boxplot by multiple factors:
p<- ggplot(data = df, aes(x = factor(delta), y = value)) +
geom_boxplot(aes(fill=factor(metric)))
The output is:
So far so good, but if I do:
p+ geom_point(aes(color = factor(metric)))
I get:
I do not know what it is doing. My goal is to color the outliers as it is done here. Note that this solution changes the inside color of the boxes to white and set the border to different colors. I want to keep the same color of the boxes while having the outliers inherit those colors. I want to know how to make the outliers get the same colors from their respective boxplots.
Do you want just to change the outliers' colour ? If so, you can do it easily by drawing boxplot twice.
p <- ggplot(data = df, aes(x = factor(delta), y = value)) +
geom_boxplot(aes(colour=factor(metric))) +
geom_boxplot(aes(fill=factor(metric)), outlier.colour = NA)
# outlier.shape = 21 # if you want a boarder
[EDITED]
colss <- c(P="firebrick3",R="skyblue", C="mediumseagreen")
p + scale_colour_manual(values = colss) + # outliers colours
scale_fill_manual(values = colss) # boxes colours
# the development version (2.1.0.9001)'s geom_boxplot() has an argument outlier.fill,
# so I guess under code would return the similar output in the near future.
p2 <- ggplot(data = df, aes(x = factor(delta), y = value)) +
geom_boxplot(aes(fill=factor(metric)), outlier.shape = 21, outlier.colour = NA)
Maybe this:
ggplot(data = df, aes(x = as.factor(delta), y = value,fill=as.factor(metric))) +
geom_boxplot(outlier.size = 1)+ geom_point(pch = 21,position=position_jitterdodge(jitter.width=0))
Related
I would like to map a variable to a continuous color scale, but if the value is a below a certain threshold, it should have a specific color. In this example, I would like to make all car models with an mpg < 20 filled with gray, and everything above with the continuous color scale:
library(ggplot2)
dat <- head(mtcars)
dat$model <- head(colnames(mtcars))
dat$is_low <- ifelse(dat$mpg < 20, TRUE, FALSE)
ggplot(dat, aes(x = model, y = mpg, fill = mpg)) +
geom_col()
Is that even possible? The closest I have seen is this solution but it maps a second variably to scale_color and not to scale_fill_continuous
Not the most elegant, but you could add a layer that's filtered to just the ones you want gray:
ggplot(dat, aes(x = model, y = mpg, fill = mpg)) +
geom_col() +
geom_col(data = dat[dat$is_low,], fill = "gray70")
Add a second layer of geom_col() where you either color each bar grey or skip if it doesn't meet your criteria. You may also want to change the limits of your coloring scale for the first layer.
ggplot(dat, aes(x = model, y = mpg, fill = mpg)) +
geom_col() + geom_col(fill=ifelse(dat$mpg < 20, 'gray', NA))
I have a boxplot with multiple groups in R.
When i add the dots within the boxplots, they are not in the center.
Since each week has a different number of boxplots, the dots are not centered within the box.
The problem is in the geom_point part.
I uploaded my data of df.m in a text file and a figure of what i get.
I am using ggplot, and here is my code:
setwd("/home/usuario")
dput("df.m")
df.m = read.table("df.m.txt")
df.m$variable <- as.factor(df.m$variable)
give.n = function(elita){
return(c(y = median(elita)*-0.1, label = length(elita)))
}
p = ggplot(data = df.m, aes(x=variable, y=value))
p = p + geom_boxplot(aes(fill = Label))
p = p + geom_point(aes(fill = Label), shape = 21,
position = position_jitterdodge(jitter.width = 0))
p = p + stat_summary(fun.data = give.n, geom = "text", fun.y = median)
p
Here is my data in a text file:
https://drive.google.com/file/d/1kpMx7Ao01bAol5eUC6BZUiulLBKV_rtH/view?usp=sharing
Only in variable 12 is in the center, because there are 3 groups (the maximum of possibilities!
I would also like to show the counting of observations. If I use the code shown, I can only get the number of observations for all the groups. I would like to add the counting for EACH GROUP.
Thank you in advance
enter image description here
Here's a solution using boxplot and dotplot and an example dataset:
library(tidyverse)
# example data
dt <- data.frame(week = c(1,1,1,1,1,1,1,1,1,
2,2,2,2,2,2,2,2,2),
value = c(6.40,6.75,6.11,6.33,5.50,5.40,5.83,4.57,5.80,
6.00,6.11,6.40,7.00,3,5.44,6.00,5,6.00),
donor_type = c("A","A","A","A","CB","CB","CB","CB","CB",
"CB","CB","CB","CB","CB","A","A","A","A"))
# create the plot
ggplot(dt, aes(x = factor(week), y = value, fill = donor_type)) +
geom_boxplot() +
geom_dotplot(binaxis='y', stackdir='center', position = position_dodge(0.75))
You should be able to adjust my code to your real dataset easily.
Edited answer with OP's dataset:
Using some generated data and geom_point():
library(tidyverse)
df.m <- df.m %>%
mutate(variable = as.factor(variable)) %>%
filter(!is.na(value))
ggplot(df.m, aes(x = variable, y = value, fill = Label)) +
geom_boxplot() +
geom_point(shape = 21, position = position_jitterdodge(jitter.width = 0)) +
scale_x_discrete("variable", drop = FALSE)
I have got a data frame with several 1000 rows in the form of
group = c("gr1","gr1","gr1","gr1","gr1","gr1","gr1","gr1","gr1","gr1","gr2","gr2","gr2","gr2","gr2","gr2","gr2","gr2","gr2","gr2","gr3","gr3","gr3","gr3","gr3","gr3","gr3","gr3","gr3","gr3")
pos = c(1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10)
color = c(2,2,2,2,3,3,2,2,3,2,1,2,2,2,1,1,1,1,1,1,2,2,2,2,2,2,1,1,2,2)
df = data.frame(group, pos, color)
and would like to make a kind of heatmap in which one axes has a continuous scale (position). The color column is categorical. However due to the large amount of data points I want to use binning, i.e. use it as a continuous variable.
This is more or less how the plot should look like:
I can't think of a way to create such a plot using ggplot2/R. I have tried several geometries, e.g. geom_point()
ggplot(data=df, aes(x=strain, y=pos, color=color)) +
geom_point() +
scale_colour_gradientn(colors=c("yellow", "black", "orange"))
Thanks for your help in advance.
Does this help you?
library(ggplot2)
group = c("gr1","gr1","gr1","gr1","gr1","gr1","gr1","gr1","gr1","gr1","gr2","gr2","gr2","gr2","gr2","gr2","gr2","gr2","gr2","gr2","gr3","gr3","gr3","gr3","gr3","gr3","gr3","gr3","gr3","gr3")
pos = c(1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10)
color = c(2,2,2,2,3,3,2,2,3,2,1,2,2,2,1,1,1,1,1,1,2,2,2,2,2,2,1,1,2,2)
df = data.frame(group, pos, color)
ggplot(data = df, aes(x = group, y = pos)) + geom_tile(aes(fill = color))
Looks like this
Improved version with 3 color gradient if you like
library(scales)
ggplot(data = df, aes(x = group, y = pos)) + geom_tile(aes(fill = color))+ scale_fill_gradientn(colours=c("orange","black","yellow"),values=rescale(c(1, 2, 3)),guide="colorbar")
I have the following example.
require(ggplot2)
# Example Data
x <- data.frame(var1=rnorm(800,0,1),
var2=rnorm(800,0,1),
var3=rnorm(800,0,1),
type=factor(rep(c("x", "y"), length.out=800)),
set=factor(rep(c("A","B","C","D"), each=200))
)
Now, I would like to plot (thin) parallel coordinate plots of these lines, with points for each of the variable values. I would like to overlay a boxplot (each of a different color for each method) on these parallel coordinate plots at the variables values. On top of this, I would like to facet for the groups and types, say using set~type. Is this possible to do using ggplot2?
Any suggestions? Thanks!
You need to put data in long format first. I didn't put in points, since the graph is already cluttered enough, but you can do so by adding a geom_point.
require(tidyr)
x$id <- 1:nrow(x)
x2 <- gather(x, var, value, var1:var3)
Boxplots
ggplot(x2, aes(var, value)) +
geom_line(aes(group = id), size = 0.05, alpha = 0.3) +
geom_boxplot(aes(fill = var), alpha = 0.5) +
facet_grid(set ~ type) +
theme_bw()
Or perhaps violins
Replacing the boxplots with violins looks pretty cool as well.
ggplot(x2, aes(var, value)) +
geom_line(aes(group = id), size = 0.05, alpha = 0.3) +
geom_violin(aes(fill = var), col = NA, alpha = 0.6) +
facet_grid(set ~ type) +
theme_bw()
I have troubles adding linear regression lines to my ggplots.
This is how it should look like:
This is how it currently looks like:
This is my code:
p <- ggplot(data = wage, aes(x = educ, y = lwage, colour = black,
cex = IQ, pch = married, alpha = 0.7)) + geom_jitter()
p1 <- p + facet_grid(urban~experclass) + geom_smooth(se=F,method="lm")
p1 + labs(x = "Education (year)", y = "Log Wage", shape = "Marital status",
colour = "Ethnicity") + guides(alpha = FALSE)
Is the position of my geom_smooth wrong? What I want is only one black regression line for each element of the plot - and not one by layer.
Furthermore what happens when I add a regression line is that the legend symbols change. Especially the IQ legend looks pretty weird. Is there something I did not consider here?
How it should look:
How it looks:
I can try to answer at least one part of your question - which is the part about plotting one regression line instead of two per panel. I don't have your data so I can't fully replicate your problem, but I think this will work.
The aesthetics in your original ggplot() call will be inherited by all the subsequent layers, including the geom_smooth.
What you seem to want is the color aesthetic (which happens to be a grouping identifier) to apply only to the jittered points and not to the line. So you can write your code like this:
p <- ggplot(data = wage, aes(x = educ, y = lwage,
cex = IQ, pch = married, alpha = 0.7)) +
geom_jitter()
p1 <- p + facet_grid(urban~experclass) +
geom_smooth(se=F,method="lm",
aes(colour = black))
or, alternatively, as one single ggplot call in a modified style:
p3 <- ggplot(data = wage,
aes(x = educ, y = lwage,
size = IQ, shape = married, alpha = 0.7)) +
geom_jitter() +
geom_smooth(se=F,method="lm",
aes(colour = black))+
facet_grid(urban~experclass)
p3