R in latex

Mix source code and documentation together
A
Write R code in LTEX using knitr, xtable and RStudio

Julyan Arbel

CREST-INSEE, Universit´ Paris-Dauphine
e

February 28, 2013

Julyan Arbel (CREST-INSEE) R code in L EX
AT February 28, 2013 1 / 10

Goal

To have a single document which includes source code for easy update.
A
Use a single software: R interfaces LTEX, as it does with C++, JAGS,
(Word, Excel?), etc.

AT February 28, 2013 2 / 10

First things

We will use the knitr package which allows you to embed R code and ﬁgures
A
in LTEX documents (it is an evolution of Sweave). See the package
homepage http://yihui.name/knitr/
A
You need a valid LTEX distribution
We will use knitr in RStudio, because it’s well integrated in it

install.packages("knitr")

library("knitr")

AT February 28, 2013 3 / 10

Write a ﬁrst document

Open a new ”R Sweave” document in RStudio
You can check that the toolbar now includes Format and Compile PDF
buttons

documentclass{article}
begin{document}

end{document}

Write your text, and insert
code chunks [Ctrl+Alt+I] for graphs or tables. A code chunk consists in R
code inside the following lines (mind to write both on single lines, with no
comment)
<<>>=
@
code values in the text with
Compile [Ctrl+Shift+I]

AT February 28, 2013 4 / 10

What to put in code chunks

Any code that you want to evaluate / not to evaluate, or show / hide: you can
either show the code, or its result, or both.
Functions
Graphs
Tables
Global options
Set working directory, etc.

AT February 28, 2013 5 / 10

Number of CRAN Packages

100
200
300
400
500
600
800
1200
1500
2000

2001−06−21 1.3

Julyan Arbel (CREST-INSEE)
2001−12−17 1.4

2002−06−12 1.5

2003−05−27 1.7

2003−11−16 1.8

2004−06−05 1.9
2004−10−12 2
Number of R packages (lien)

2005−06−18 2.1

AT
R code in L EX
2005−12−16 2.2
R Version

2006−05−31 2.3

2006−12−12 2.4
2007−04−12 2.5

2007−11−16 2.6
2008−03−18 2.7

2008−10−18 2.8

2009−09−17 2.9
110
129
162
219
273
357
406
548
647
739
911
1000
1300
1427
1614
1952

February 28, 2013
6 / 10

Number of R packages: now the code

rv <- seq(1.3, 2.9, .1)
pckg.num <- c(110,129,162,219,273,357,406,548,647,739,911,1000,1300,
rv.dates <- c("2001-6-21", "2001-12-17","2002-06-12","2003-05-27",
"2003-11-16","2004-06-05","2004-10-12","2005-06-18","2
"2006-12-12","2007-04-12","2007-11-16","2008-03-18","2
pckg.fit <- lm(pckg.num~rv)
par(mar=c(7, 5, 5, 3), las=2)
plot(as.POSIXct(rv.dates), pckg.num, xlab="",ylab="",col="red", log=
axis.POSIXct(1, 1:16, rv.dates, format="%Y-%m-%d")
axis(2, at=c(100,200,300,400,500,600,800,100,1200,1500,2000))
mtext("Number of CRAN Packages", side=2, line=3, las=3)
axis.POSIXct(3, rv.dates, rv.dates, labels=as.character(rv))
mtext("R Version", side=3, line=3, las=1)
axis(4, pckg.num)
abline(v=as.POSIXct(rv.dates), col="lightgray", lty="dashed")
abline(h=pckg.num, col="lightgray", lty="dashed")
abline(lm(log10(pckg.num)~as.POSIXct(rv.dates)), col="red")

AT February 28, 2013 7 / 10

To show graphs

Use that kind of code chunk
<<echo=FALSE,out.width=’.7textwidth’,fig=TRUE,include=TRUE>>=
library(ggplot2)
qplot(speed, dist, data=cars)+geom_smooth()
@

AT February 28, 2013 8 / 10

To show graphs: result

100

dist

50

0

5 10 15 20 25
speed

AT February 28, 2013 9 / 10

To show data tables

Use that kind of code chunk
<<echo=FALSE,results=asis>>=
load(file="data.Rdata")
row.names(data) = c(...)
library(xtable)
tab=xtable(data)
digits(tab)=1
print(tab,floating=FALSE,include.rownames=TRUE,type="latex")
@

AT February 28, 2013 10 / 10

R in latex

More Related Content

What's hot (6)

More from Julyan Arbel (20)

R in latex