In an effort to help populate the R tag here, I am posting a few questions I have often received from students. I have developed my own answers to these over the years, but perhaps there are better ways floating around that I don't know about.
The question: I just ran a regression with continuous y and x but factor f (where levels(f) produces c("level1","level2"))
thelm <- lm(y~x*f,data=thedata)
Now I would like to plot the predicted values of y by x broken down by groups defined by f. All of the plots I get are ugly and show too many lines.
My answer: Try the predict() function.
##restrict prediction to the valid data
##from the model by using thelm$model rather than thedata
thedata$yhat <- predict(thelm,
newdata=expand.grid(x=range(thelm$model$x),
f=levels(thelm$model$f)))
plot(yhat~x,data=thethedata,subset=f=="level1")
lines(yhat~x,data=thedata,subset=f=="level2")
Are there other ideas out there that are (1) easier to understand for a newcomer and/or (2) better from some other perspective?
The effects package has good ploting methods for visualizing the predicted values of regressions.
thedata<-data.frame(x=rnorm(20),f=rep(c("level1","level2"),10))
thedata$y<-rnorm(20,,3)+thedata$x*(as.numeric(thedata$f)-1)
library(effects)
model.lm <- lm(formula=y ~ x*f,data=thedata)
plot(effect(term="x:f",mod=model.lm,default.levels=20),multiline=TRUE)
Huh - still trying to wrap my brain around expand.grid(). Just for comparison's sake, this is how I'd do it (using ggplot2):
thedata <- data.frame(predict(thelm), thelm$model$x, thelm$model$f)
ggplot(thedata, aes(x = x, y = yhat, group = f, color = f)) + geom_line()
The ggplot() logic is pretty intuitive, I think - group and color the lines by f. With increasing numbers of groups, not having to specify a layer for each is increasingly helpful.
I am no expert in R. But I use:
xyplot(y ~ x, groups= f, data= Dat, type= c('p','r'),
grid= T, lwd= 3, auto.key= T,)
This is also an option:
interaction.plot(f,x,y, type="b", col=c(1:3),
leg.bty="0", leg.bg="beige", lwd=1, pch=c(18,24),
xlab="",
ylab="",
trace.label="",
main="Interaction Plot")
Here is a small change to the excellent suggestion by Matt and a solution similar to Helgi but with ggplot. Only difference from above is that I have used the geom_smooth(method='lm) which plots regression lines directly.
set.seed(1)
y = runif(100,1,10)
x = runif(100,1,10)
f = rep(c('level 1','level 2'),50)
thedata = data.frame(x,y,f)
library(ggplot2)
ggplot(thedata,aes(x=x,y=y,color=f))+geom_smooth(method='lm',se=F)
Related
I want to compare the fit of different distributions to my data in a single plot. The qqcomp function from the fitdistrplus package pretty much does exactly what I want to do. The only problem I have however, is that it's mostly written using base R plot and all my other plots are written in ggplot2. I basically just want to customize the qqcomp plots to look like they have been made in ggplot2.
From the documentation (https://www.rdocumentation.org/packages/fitdistrplus/versions/1.0-14/topics/graphcomp) I get that this is totally possible by setting plotstyle="ggplot". If I do this however, no points are showing up on the plot, even though it worked perfectly without the plotstyle argument. Here is a little example to visualize my problem:
library(fitdistrplus)
library(ggplot2)
set.seed(42)
vec <- rgamma(100, shape=2)
fit.norm <- fitdist(vec, "norm")
fit.gamma <- fitdist(vec, "gamma")
fit.weibull <- fitdist(vec, "weibull")
model.list <- list(fit.norm, fit.gamma, fit.weibull)
qqcomp(model.list)
This gives the following output:
While this:
qqcomp(model.list, plotstyle="ggplot")
gives the following output:
Why are the points not showing up? Am I doing something wrong here or is this a bug?
EDIT:
So I haven't figured out why this doesn't work, but there is a pretty easy workaround. The function call qqcomp(model.list, plotstyle="ggplot") still returns an ggplot object, which includes the data used to make the plot. Using that data one can easily write an own plot function that does exactly what one wants. It's not very elegant, but until someone finds out why it's not working as expected I will just use this method.
I was able to reproduce your error and indeed, it's really intriguing. Maybe, you should contact developpers of this package to mention this bug.
Otherwise, if you want to reproduce this qqplot using ggplot and stat_qq, passing the corresponding distribution function and the parameters associated (stored in $estimate):
library(ggplot2)
df = data.frame(vec)
ggplot(df, aes(sample = vec))+
stat_qq(distribution = qgamma, dparams = as.list(fit.gamma$estimate), color = "green")+
stat_qq(distribution = qnorm, dparams = as.list(fit.norm$estimate), color = "red")+
stat_qq(distribution = qweibull, dparams = as.list(fit.weibull$estimate), color = "blue")+
geom_abline(slope = 1, color = "black")+
labs(title = "Q-Q Plots", x = "Theoritical quantiles", y = "Empirical quantiles")
Hope it will help you.
I'm trying to add second order curve to scatterplot.
I've read the answers to previous similar questions and here's what I came up with:
x<-log2(c(100,500,1000,2000,4000))
y<-c(3.6,1.308,1.065,.960,.908)
plot(x,y,pch=1)
mod_<-lm(y~poly(x,2,raw=TRUE))
lines(x,predict(mod_),col='red',lty=2)
Still, I get linear segments instead of smooth curve.
What mistake am I not seeing here ? Thanks !
You are calling predict by passing the model only. This only results in the model being evaluated at the values you specified in your lm call (that is x).
You need to supply a new set of values at which the model will be evaluated.
For, instance, this gives you a nice smooth line:
x<-log2(c(100,500,1000,2000,4000))
y<-c(3.6,1.308,1.065,.960,.908)
plot(x,y,pch=1)
mod_<-lm(y~poly(x,2,raw=TRUE))
# Define the new points at which you want to evaluate your model
new.x <- seq(6, 12, 0.1)
lines(new.x, predict(mod_, newdata = list(x=new.x)),col='red',lty=2)
You can also use ggplot2 like this
library(ggplot2)
df <- data.frame(x, y)
ggplot(data=df, aes(x, y))+geom_point()+stat_smooth(method="lm", formula = y ~ poly(x, 2, raw =TRUE))
I apologize first for bringing what I imagine to be a ridiculously simple problem here, but I have been unable to glean from the help file for package 'polynom' how to solve this problem. For one out of several years, I have two vectors of x (d for day of year) and y (e for an index of egg production) data:
d=c(169,176,183,190,197,204,211,218,225,232,239,246)
e=c(0,0,0.006839425,0.027323127,0.024666883,0.005603878,0.016599262,0.002810977,0.00560387 8,0,0.002810977,0.002810977)
I want to, for each year, use the poly.calc function to create a polynomial function that I can use to interpolate the timing of maximum egg production. I want then to superimpose the function on a plot of the data. To begin, I have no problem with the poly.calc function:
egg1996<-poly.calc(d,e)
egg1996
3216904000 - 173356400*x + 4239900*x^2 - 62124.17*x^3 + 605.9178*x^4 - 4.13053*x^5 +
0.02008226*x^6 - 6.963636e-05*x^7 + 1.687736e-07*x^8
I can then simply
plot(d,e)
But when I try to use the lines function to superimpose the function on the plot, I get confused. The help file states that the output of poly.calc is an object of class polynomial, and so I assume that "egg1996" will be the "x" in:
lines(x, len = 100, xlim = NULL, ylim = NULL, ...)
But I cannot seem to, based on the example listed:
lines (poly.calc( 2:4), lty = 2)
Or based on the arguments:
x an object of class "polynomial".
len size of vector at which evaluations are to be made.
xlim, ylim the range of x and y values with sensible defaults
Come up with a command that successfully graphs the polynomial "egg1996" onto the raw data.
I understand that this question is beneath you folks, but I would be very grateful for a little help. Many thanks.
I don't work with the polynom package, but the resultant data set is on a completely different scale (both X & Y axes) than the first plot() call. If you don't mind having it in two separate panels, this provides both plots for comparison:
library(polynom)
d <- c(169,176,183,190,197,204,211,218,225,232,239,246)
e <- c(0,0,0.006839425,0.027323127,0.024666883,0.005603878,
0.016599262,0.002810977,0.005603878,0,0.002810977,0.002810977)
egg1996 <- poly.calc(d,e)
par(mfrow=c(1,2))
plot(d, e)
plot(egg1996)
I am very, very new to R so please forgive the basic nature of my question. In short, I have done a lot of Google searching to try to answer this, but I find that even the basic guides available, and simple discussions on forums are assuming more prior knowledge than I have, especially when it comes to outlining what all of the coding terms are and what changing them means for a plot.
In short I have a tab formatted table with three columns of data that I wish to plot densities for on a single graph. I would like the lines to be different patterns (dotted, dashed etc. whatever makes it easy to tell them apart, I cannot use colours as my supervisor is colour blind).
I have code that reads in the data and makes accessible the columns I am interested in:
mydata <- read.table("c:/Users/Demon/Desktop/Thesis/Fst_all_genome.txt", header=TRUE,
sep="\t")
fstdata <- data.frame(Fst_ceu_mkk =rnorm(10),
Fst_ceu_yri =rnorm(10),
Fst_mkk_yri =rnorm(10))
Where do I go from here?
Appendix A of 'An Introduction to R' has a nice walkthrough tutorial you can do in ten minutes; it teaches among other things about line types etc
After that, plotting densities was explained dozens of times here too; search in the search box above for eg '[r] density'. There is also the R Graph Gallery (possibly down right now) and more.
A nice, free guide I often recommend is John Verzani's simpleR which stresses graphs a lot and will teach you what you need here.
Two options for you to explore using high-level graphics.
# dummy data
d = data.frame(x = rnorm(10), y = rnorm(10), z = rnorm(10))
You first need to reshape the data from wide to long format,
require(reshape2)
m = melt(d)
ggplot2 graphics
require(ggplot2)
ggplot(data = m, mapping = aes(x = value, linetype = variable)) +
geom_line(stat = "density")
Lattice graphics
Using the same melt()ed data,
require(lattice)
densityplot( ~ value, data = m, group = variable,
auto.key = TRUE, par.settings = col.whitebg())
If you need something very simple, you could do simply:
plot(density(mydata$col_1))
lines(density(mydata$col_2), lty = 2)
lines(density(mydata$col_2), lty = 3)
If the second and third density curves are far away from the first, you'll need define xy limits of the plotting region explicitly:
dens1 <- density(mydata$col_1)
dens2 <- density(mydata$col_2)
dens3 <- density(mydata$col_3)
plot(dens1, xlim = range(dens1$x, dens2$x, dens3$x),
ylim = range(dens1$y, dens2$y, dens3$y))
lines(density(mydata$col_2), lty = 2)
lines(density(mydata$col_2), lty = 3)
Hope this helps.
I just came by the following plot:
And wondered how can it be done in R? (or other softwares)
Update 10.03.11: Thank you everyone who participated in answering this question - you gave wonderful solutions! I've compiled all the solution presented here (as well as some others I've came by online) in a post on my blog.
Make.Funny.Plot does more or less what I think it should do. To be adapted according to your own needs, and might be optimized a bit, but this should be a nice start.
Make.Funny.Plot <- function(x){
unique.vals <- length(unique(x))
N <- length(x)
N.val <- min(N/20,unique.vals)
if(unique.vals>N.val){
x <- ave(x,cut(x,N.val),FUN=min)
x <- signif(x,4)
}
# construct the outline of the plot
outline <- as.vector(table(x))
outline <- outline/max(outline)
# determine some correction to make the V shape,
# based on the range
y.corr <- diff(range(x))*0.05
# Get the unique values
yval <- sort(unique(x))
plot(c(-1,1),c(min(yval),max(yval)),
type="n",xaxt="n",xlab="")
for(i in 1:length(yval)){
n <- sum(x==yval[i])
x.plot <- seq(-outline[i],outline[i],length=n)
y.plot <- yval[i]+abs(x.plot)*y.corr
points(x.plot,y.plot,pch=19,cex=0.5)
}
}
N <- 500
x <- rpois(N,4)+abs(rnorm(N))
Make.Funny.Plot(x)
EDIT : corrected so it always works.
I recently came upon the beeswarm package, that bears some similarity.
The bee swarm plot is a
one-dimensional scatter plot like
"stripchart", but with closely-packed,
non-overlapping points.
Here's an example:
library(beeswarm)
beeswarm(time_survival ~ event_survival, data = breast,
method = 'smile',
pch = 16, pwcol = as.numeric(ER),
xlab = '', ylab = 'Follow-up time (months)',
labels = c('Censored', 'Metastasis'))
legend('topright', legend = levels(breast$ER),
title = 'ER', pch = 16, col = 1:2)
(source: eklund at www.cbs.dtu.dk)
I have come up with the code similar to Joris, still I think this is more than a stem plot; here I mean that they y value in each series is a absolute value of a distance to the in-bin mean, and x value is more about whether the value is lower or higher than mean.
Example code (sometimes throws warnings but works):
px<-function(x,N=40,...){
x<-sort(x);
#Cutting in bins
cut(x,N)->p;
#Calculate the means over bins
sapply(levels(p),function(i) mean(x[p==i]))->meansl;
means<-meansl[p];
#Calculate the mins over bins
sapply(levels(p),function(i) min(x[p==i]))->minl;
mins<-minl[p];
#Each dot is one value.
#X is an order of a value inside bin, moved so that the values lower than bin mean go below 0
X<-rep(0,length(x));
for(e in levels(p)) X[p==e]<-(1:sum(p==e))-1-sum((x-means)[p==e]<0);
#Y is a bin minum + absolute value of a difference between value and its bin mean
plot(X,mins+abs(x-means),pch=19,cex=0.5,...);
}
Try the vioplot package:
library(vioplot)
vioplot(rnorm(100))
(with awful default color ;-)
There is also wvioplot() in the wvioplot package, for weighted violin plot, and beanplot, which combines violin and rug plots. They are also available through the lattice package, see ?panel.violin.
Since this hasn't been mentioned yet, there is also ggbeeswarm as a relatively new R package based on ggplot2.
Which adds another geom to ggplot to be used instead of geom_jitter or the like.
In particular geom_quasirandom (see second example below) produces really good results and I have in fact adapted it as default plot.
Noteworthy is also the package vipor (VIolin POints in R) which produces plots using the standard R graphics and is in fact also used by ggbeeswarm behind the scenes.
set.seed(12345)
install.packages('ggbeeswarm')
library(ggplot2)
library(ggbeeswarm)
ggplot(iris,aes(Species, Sepal.Length)) + geom_beeswarm()
ggplot(iris,aes(Species, Sepal.Length)) + geom_quasirandom()
#compare to jitter
ggplot(iris,aes(Species, Sepal.Length)) + geom_jitter()