The following few commands demonstrate the basics of using R.
To create vectors, matrices and data frames
## creates two vectors x, y n <- 30 x <- 1:n # a vector of (1,2,...,n) x # print x on screen x[1:5] # print first 5 elements x1 <- seq(from=0, to=1, by=0.1) # a sequence with equal distance 0.1 x2 <- runif(n, min=0, max=1) # realisations of U(0,1) rvs x3 <- rbinom(n, size=1, prob=0.5) # realisations of B(1,0.5) rvs y <- rnorm(n, mean=0, sd=1) # realisations of N(0,1) rvs xx1 <- cbind(x, y) # create a matrix by merging by column dim(xx1) # dimension of matrix xx2 <- rbind(x2, y) # create a matrix by merging by row dim(xx2) DF <- data.frame(seq=x, unif=x2, norm=y, binary=x3) DF # data frame containing 4 variables names(DF) # names of variables in data frame DF$seq # access data frame contents DF$norm
Summary statistics
mean(y) # sample mean var(y) # sample variance sd(y) # sample standard deviation range(y) # min and max median(y) # median value quantile(y, prob=0.25) # 25%-ile summary(y) # a few summary statistics of vector summary(DF) # summary stats of each variable in data frame table(x3) # table of x3 table(rpois(n, lambda=1), x3) # cross-table of Poi(1) samples and x3
Draw a few simple plots
## plots plot(x, y) # scatter plot of x and y plot(xx1) par(mfrow = c(2, 1)) # arrange multiple plots on one page plot(x, y) lines(x, y, col="red") # add lines plot(x2, y) lines(x2, y, col="blue") hist(x2, prob=TRUE) # histogram of x2 with probs not freqs hist(y,col="yellow",20) # histogram of y with 20 breaks boxplot(x2) # boxplot of x2 boxplot(norm~binary, data=DF) # boxplots of norm, split by bin x11()ΨΨΨΨΨΨΨΨΨΨΨ # open new plotting window barplot(table(x3)) # Barplot of x3
R as a calculator
## mathematical functions x <- 1:10 x*x # elementwise multiplication 2^x # elementwise power x^2 # elementwise square exp(x) # elementwise exponential log(x) # nb: log(0) = -Inf x + sin(x*pi/4) # nb: pi = 3.41593 ## square matrix and inverse A <- matrix(c(2,1,-1,3), nrow=2, ncol=2) # read column-wise A A[1,]ΨΨΨΨΨΨΨ# row 1 A[,2]ΨΨΨΨΨΨΨ# column 2 A[2,1]ΨΨΨΨΨΨ# element at row 2 and column 1 t(A) # transpose of A solve(A) # inverse of A B <- matrix(c(1,2,1,2), nrow=2, ncol=2) solve(B) # non-invertible, matrix is singular det(B)ΨΨΨΨΨΨ# determinant of B ## matrix multiplication A%*%B
Using for loops
Repeating operations can be made simple by using for loops.
out <- matrix(0, nrow=20, ncol=2) # storage
for (i in 1:20){ # define loop variable and set of values to take
a=2*i
b=3+i
out[i,1] <- a
out[i,2] <- b
# out[i,] <- c(a,b) # store as a row vector each time
}
out
How to find help
To access the help page,
help.start() # opens webpage, click on "An Introduction to R" ?runif # opens manual page for simulating uniform rvs help(lm) # opens manual page for linear regression
Installing and loading libraries
By default, R contains a limited number of functions for common need. There are 100s of packages available on the Comprehensive R Archive Network (CRAN) that contain functions for specialist use.
install.packages("censReg") # install package
# select a local CRAN mirror (UK, Bristol)
library("censReg") # load in library
Accessing data in R
There are some example data sets available in the base package of R.
data() # lists data sets available data(women) # makes women available ?women # information about the data names(women) # name of the variables summary(women) # data frame summary attach(women) # adds women to the search list summary(height) # The same variable now available by name detach(women) # remove women to the search list summary(height) # comprehend?
Loading data from file.
There are a few ways to load data into R depending on the file format.
load("DATA.Rdata") # file type .Rdata
read.table("DATA.txt") # file type .txt
read.table("DATA.dat") # file type .dat
read.csv("DATA.csv") # file type .csv
Remember to change your working directory using setwd("PATH") or put the full path to the data file in the above commands.
Memory management
ls() # list items in memory a <- "NEW OBJECT" # create new character object ls() # object "a" appears in list rm(a)ΨΨΨΨΨΨΨΨ # remove "a" from memory ls()