Introduction to R - 1st lesson (introduction)

• is a free statistical software
• can be installed on Linux, Windows and Mac OS X
• has a number of additional packages providing a wide number of functions for statistics and scientific programming

R can be downloaded at http://cran.univ-paris1.fr with different installation instructions depending on your OS

For Ubuntu user, R can be installed with the command line

sudo apt-get install r-base-core r-base-dev

(see also

http://tuxette.nathalievialaneix.eu/?p=1380&lang=en

for a more up-to-date installation)

• On Windows and Mac OS X, R is started by double clicking on its icon. A very basic GUI then opens up

• On Linux/Unix, R is started using the command

  R
  

  in a terminal and the program starts with no GUI

RStudio is a set of integrated tools designed to help you use R. It includes a GUI:

RStudio is a set of integrated tools designed to help you use R. It includes a GUI.

Good practices with your statistical analyses:

• systematically create and save a R script that contains all the command lines that you have used (plain text file named XXX.R)
• comment your scripts using the character #

# this is a comment in a R script
data(iris)

• you can save your results as well (R data files are named XXX.rda)

Additional packages can be installed with

• the menu “Install packages” on Windows and Mac OS X
• the command line install.packages("RColorBrewer") where RColorBrewer is the name of the package that you want to install
• the menu “Tools/Install packages” in RStudio

In all cases, a menu pops up asking you to choose your CRAN repository: select Paris 2 which is the repository hosted by University Paris 1!

Once a package is installed, you can load it using the command line

library(RColorBrewer)

and all its functionalities are available until you close your R session.

• Help on a given function help(mean)
• Help on a given topic help.search("mean") returns a list of functions that correspond to this topic
• Demonstrations: the command line demo() lists available demonstrations. Try demo(graphics) to have a demonstration of the graphical capabilities of R

• official R manuals http://cran.univ-paris1.fr/manuals.html (all in English)
• non official R documentation http://cran.univ-paris1.fr/other-docs.html (some of them are in French) and, in particular, the short http://cran.univ-paris1.fr/doc/contrib/Kauffmann_aide_memoire_R.pdf (in French) that all of you must have with him/her (5 pages)
• the forum stackoverflow and its section dedicated to R: http://stackoverflow.com/questions/tagged/r
• the MOOC “Computing for data analysis” https://www.coursera.org/course/compdata

R can be used for basic operations:

log((1 + sqrt(3)/2)^4)

[1] 2.495243

See operators: +, -, *, /, etc

See functions: log, exp, cos, sin, tan, etc

Results can be stored in a variable

a <- cos(pi/4)
a

[1] 0.7071068

b = sin(pi/3+1)
b

[1] 0.888651

The mode of a variable indicates how it is stored in memory:

mode(a)

[1] "numeric"

is.numeric(a)

[1] TRUE

Numeric variables can also be specifically integers:

b <- 1L; mode(b)

[1] "numeric"

is.integer(b); is.integer(a); is.numeric(b)

[1] TRUE

[1] FALSE

[1] TRUE

b <- a; b; mode(b)

[1] 0.7071068

[1] "numeric"

b <- "a"; b; mode(b)

[1] "a"

[1] "character"

b <- as.character(a); b; mode(b)

[1] "0.707106781186548"

[1] "character"

b <- as.numeric(b); b; mode(b)

[1] 0.7071068

[1] "numeric"

b <- as.numeric("a"); b;

Warning: NAs introduced by coercion

[1] NA

In R missing values are coded with NA:

d <- NA; d; is.na(d)

[1] NA

[1] TRUE

b <- "ABCABD"; nchar(b)

[1] 6

substr(b, 1, 3)

[1] "ABC"

b <- tolower(b); b; toupper(b)

[1] "abcabd"

[1] "ABCABD"

a <- "a"; b <- "abcabd"; paste(a,b,sep="-")

[1] "a-abcabd"

strsplit(b,a)

[[1]]
[1] ""   "bc" "bd"

grep(a, b); gsub(a,"",b)

[1] 1

[1] "bcbd"

d <- is.numeric(b); d; mode(d)

[1] FALSE

[1] "logical"

as.numeric(d); as.character(d); as.logical(1)

[1] 0

[1] "FALSE"

[1] TRUE

e <- !d; e

[1] TRUE

d|e; e&d

[1] TRUE

[1] FALSE

e

[1] TRUE

e+e;

[1] 2

as.logical("0")

[1] NA

Vectors can be used to concatenate several variables of the same mode

a.vector <- c(a, b, d, e); a.vector

[1] "a"      "abcabd" "FALSE"  "TRUE"  

mode(a.vector); length(a.vector)

[1] "character"

[1] 4

nchar(a.vector)

[1] 1 6 5 4

a <- c(1:5,NA); b <- seq(1,100,length=5); a; b

[1]  1  2  3  4  5 NA

[1]   1.00  25.75  50.50  75.25 100.00

a*b # !! this is not a dot product

[1]   1.0  51.5 151.5 301.0 500.0    NA

a; sum(a); sum(a, na.rm=TRUE)

[1]  1  2  3  4  5 NA

[1] NA

[1] 15

sum(a > 2, na.rm=TRUE)

[1] 3

a; b <- rep(1:6, each=2); b

[1]  1  2  3  4  5 NA

 [1] 1 1 2 2 3 3 4 4 5 5 6 6

unique(b); intersect(a,b); union(a,b)

[1] 1 2 3 4 5 6

[1] 1 2 3 4 5

[1]  1  2  3  4  5 NA  6

b <- rep(1:3, 2); b; 

[1] 1 2 3 1 2 3

b>2; which(b>2); b[c(3,6)]; b[b>2]

[1] FALSE FALSE  TRUE FALSE FALSE  TRUE

[1] 3 6

[1] 3 3

[1] 3 3

b <- rep(1:3, each=2); b; 

[1] 1 1 2 2 3 3

which(b>5); length(which(b>5))

integer(0)

[1] 0

b==2; 

[1] FALSE FALSE  TRUE  TRUE FALSE FALSE

A factor is a vector with a pre-defined list of values

a.factor <- as.factor(a.vector); a.factor

[1] a      abcabd FALSE  TRUE  
Levels: a abcabd FALSE TRUE

class(a.factor); levels(a.factor); nlevels(a.factor)

[1] "factor"

[1] "a"      "abcabd" "FALSE"  "TRUE"  

[1] 4

a.factor[3] <- "2"; a.factor

[1] a      abcabd <NA>   TRUE  
Levels: a abcabd FALSE TRUE

a.factor[3] <- "a"; a.factor

[1] a      abcabd a      TRUE  
Levels: a abcabd FALSE TRUE

levels(a.factor) <-5:8; a.factor

[1] 5 6 5 8
Levels: 5 6 7 8

as.character(a.factor)

[1] "5" "6" "5" "8"

as.numeric(a.factor)

[1] 1 2 1 4

ls()

[1] "a"        "a.factor" "a.vector" "b"        "d"        "e"       
[7] "iris"    

rm(iris); ls()

[1] "a"        "a.factor" "a.vector" "b"        "d"        "e"       

rm(list=ls())

This information is also available (in a more detailed manner) in the panel “Environment” of RStudio.

a <- c(1, 5, 3, 4); b <- letters[5:7]
a; b

[1] 1 5 3 4

[1] "e" "f" "g"

How to obtain from a and b:

d

[1] "1" "5" "e" "f" "g" "3" "4"

a <- c(1, 2, NA, 3:7); a; b <- 1:7; all.equal(a,b)

[1]  1  2 NA  3  4  5  6  7

[1] "Numeric: lengths (8, 7) differ"

How to remove missing values from a such that

all.equal(a, b)

[1] TRUE

set.seed(1119)
as <- sample(1:5, 100, replace=TRUE)
head(as); tail(as)

[1] 4 3 4 1 4 1

[1] 5 2 4 5 4 5

What do the functions set.seed, sample, head and tail do? Find in one command line the number of “5” in this vector. The result must be equal to:

[1] 17