Introduction to R

Exercise 1

1. Start R . Select R from the Start -> All Programs -> R menu.

2. On the command line, type demo(graphics). Follow prompts. You might prefer to re-size the
windows.

3. On the command line, type demo(image). This demonstration is concerned with representations
of 3D data.

4. Get help on the function q, by typing ?q.

5. Quit R , by typing q().

It will be useful to create a location for files for this course. Your course account includes a network
drive, drive H:. Create a new directory on this drive, called rcourse say. Follow these instructions to
create a shortcut on the desktop for R to open using this location.

1. Right click on the desktop and select New -> Shortcut

2. Select browse in the dialogue box.

3. Find c:\Program Files\R\rw2001\bin\Rgui.exe

4. Click Next.

5. Enter a suitable name for the shortcut, and click finish.

Now, modify the properties of the shortcut as follows

1. Right click the R shortcut, and select Properties.

2. Change the Start in entry to the network directory you created, and click Ok.

Now, start R using this new shortcut. Type shell("dir"). Which directory is referred to? Finally,
quit this R session.

1
 Introduction to R
Exercise 2

Start R using the shortcut created earlier. Download the file

http://stats.ma.ic.ac.uk/das01/public_html/RCourse/hills.txt

and save it in the directory created earlier. This file contains the Scottish hill races data set. Now,
enter the following commands

1. Read the file, assigning the result to the object hills

hills <- read.table("hills.txt")

2. Examine the object. Note column and row names

hills

3. Construct a scatterplot matrix

pairs(hills)

How many variables do you think such plots are suitable for?

4. Make the columns of the hills object available by name

attach(hills)

5. Construct a scatter plot. The function call in this way means the first argument is the horizontal
axis

plot(dist,time)

6. Interact with the plot to label points - right click to finish

identify(dist,time,row.names(hills))

7. Compute a linear regression of time against distance

lm(time~dist)

8. Obtain more information about the regression

summary(lm(time ~ dist))

2
 9. Add the least squares regression line - note anonymous function call

abline(lm(time~dist))

10. Obtain some diagnostics plots - note the different arguments to the plot function. Be aware of
the prompt in the Console.

plot(lm(time~dist))

11. there are many pre-defined system objects. Display the value of pi - note that this is a reserved
word

pi

12. List objects in current working space

ls()

13. Display the object ls

ls

14. Create a copy of the hills object

hill.cp <- hills

15. List objects in current working space

ls()

16. Delete the copy. Note that there is no undelete functionality.

rm(hill.cp)

Finally, quit this session.

3
 Introduction to R
Exercise 3

This sheet is primarily concerned with vectors and arithmetic.

1. Create a vector of coefficients for a quadratic equation, using the sample function. Here, we draw
a sample of size 3 from −20, −19, . . . , 19, 20 with replacement

coeffs <- sample(-20:20,3,replace=T)

2. Determine the class of the object coeffs.

class(coeffs)

3. Determine the length of the object coeffs

length(coeffs)

4. Determine the names associated with the vector

names(coeffs)

What was the result?

5. Assign some names. Note the function call occurring on the right hand of the assignment operator.

names(coeffs) <- c("a","b","c")

What is the class of the names of the vector coeffs?

6. Prepare to plot the equation, by constructing a regularly spaced vector for the horizontal axis

x <- seq(-3,3,length=200)

7. Evaluate the quadratic at each point in vector x

y <- coeffs[1]*x^2+coeffs[2]*x+coeffs[3]

8. Construct the plot

plot(x,y)

plot(y=y,x=x)

Note the different styles of function call.

4
 9. Does the equation have real roots? Compute the discriminant

coeffs[2]^2-4*coeffs[1]*coeffs[3]

10. Oops, we didn’t retain the value! R stores the last unassigned object in the system object
.Last.value

disc <- .Last.value

11. Create a vector of type character, and display the second element

chr.vec <- sample(letters,5); chr.vec[2]

Problems
1. Compute the real roots of the quadratic equation

x2 + x + 1 = 0

Recall the formula for the roots of a quadratic

2. Without using R , determine the result of the following computation

x <- c(1,2,3)
x[1]/x[2]^3-1+2*x[3]-x[2-1]

3. Generate a regular grid between -50 and 50. Construct separate plots of log(x), exp(x), sin(x),
sin(2x), sin(x)/cos(x). Examine the cumulative sum of the final function. Experiment with the
argument type of the plot function.

5
 Introduction to R
Exercise 4

This sheet is concerned with logical operations and factors.

1. Create a logical vector

x <- seq(-3,3,length=200) > 0

2. negate this vector

!x

3. Compute the truth table for logical AND

c(T,T,F,F) & c(T,F,F,T)

4. Explore arithmetic with logical and numeric

1:3 + c(T,F,T)

5. Compute the intersection of {1, 2, . . . , 10} and {5, 6, . . . , 15}

intersect(1:10,5:15)

6. Create a factor

drinks <- factor(c("beer","beer","wine","water")

7. Examine the representation of the factor

unclass(drinks)

Problems
1. Compute the truth table for logical OR. The function R computes the logical EXCLUSIVE-OR.
What is the difference between the two?

2. Consider the vector 1:K, where K is a positive integer. Write an R command that determines how
many elements in the vector are exactly divisible by 3.

3. Write an R command to evaluate the proportion of beer in the drinks factor object.

6
 Introduction to R
Exercise 5
This sheet is concerned with data frames and matrices.

1. Examine the row and column names of the hills object

row.names(hills)
names(hills)

2. Create data frames. Note row and column names

x1.df <- data.frame(1:10,I(letters[1:10]),factor(letters[1:10]))

3. Compute the mean of column X1

mean(x2.df$X1)

4. Create a matrix

x.mat <-matrix(1:12,nrow=3,ncol=4)

5. Examine the default dimension names of the matrix

dimnames(x.mat)

6. Assign some dimnames

dimnames(x.mat) <- list(letters[1:3],letters[1:4])

7. Combine matrices

xx <- cbind(x.mat,x.mat)
xxx <- rbind(x.mat,x.mat)
rbind(xx,xxx)

8. Explore indexing

x <- 1:10
names(x) <- letters[x]
x[1:3]
x[c(1,10)]
x[c(-1,-2)]
x[ x > 5]
x[c("a","d")]
x[]
jj1 <- matrix(1:100,ncol=10)
jj1[1:5,]
jj1[1:4,x[x <3]]

7
 9. Compute row and column sums of a data frame

x <- matrix(1:10,ncol=2)
lapply(x,sum)
sapply(log(x),sum)

10. Create a list, and examine elements

x.lis <- list(a=1:10,b=letters[1:3],b=matrix(1:10,ncol=2))

11. Element-wise arithmetic with matrices

x.mat <- matrix(1:10,ncol=2)

12. Matrix multiplication

x.mat %*% t(x.mat)

13. Compute row and column sums of a matrix

apply(x,1,sum)
apply(x,2,sum)

Problems
1. Construct a 2×2 data frame, X say. Experiment with X^(1:K), where K takes values 1:4. How
does the recycling rule behave? What happens if you remove the brackets from the command?

2. The function system.time returns timings for R operations. Examine the help system about this
function. For a 107 × 2 matrix, X, and vector y of length 107 /2 compute (a number of times) X t y
using matrix multiplication and the function crossproduct. Which is quicker?

8
 Introduction to R
Exercise 6

This sheet is concerned with usage issues.

1. Determine what objects are in the current workspace

ls()
objects()

2. Create and edit a new data frame

a.df <- data.frame()

3. Create, then delete some objects. Note the multiple assignment

jj1 <- jj2 <- jj3 <- a.df

4. Examine the search path

search()
attach(hills)
search()

5. Write the hills object to a file

write.table("test.dat",hills)

Where is this file situated on the file system. Examine the default output format by viewing the
file with your favourite text editor.

6. Get some help

?mean
help.start()

7. Do a silly large computation.

sin(matrix(0,nrow=5000,ncol=5000))

Use hESCi to interrupt. Note that this might take some time to return control to the console.

Problems
1. Generate a matrix of size n × p. Use the function as.data.frame to coerce the matrix to a data
frame. Which object requires more storage space?

9
 Introduction to R
Exercise 7

This sheet is concerned with basic statistics and graphics.

1. Generate a sample of random normal deviates, and a sample of random exponential deviates.

x<- rnorm(50)
y <- rnorm(50,0,1)

2. Compute some summaries

mean(x)
sqrt(var(x))
cor(x,y)
cor(cbind(x,y))

3. Try the summary function

summary(x)
summary(cbind(x,y))

4. Let X ∼ N (0, 1) and Y ∼ Exp(2). Compute P (X > 1.644) and find q such that P (Y < q) = 0.75.

1-pnorm(1.644)
qexp(0.75,2)

5. Use the sample function to obtain a random sample of 10 realisations in a biased coin experiment

sample(c("Head","Tail"), 10,probs=c(0.3,0.7),replace=T)

6. Load the package SuppDists and examine the help

help(package="SuppDists")

7. Experiment with set.seed

set.seed(1)
runif(10)
set.seed(1)
runif(10)
runif(10)

8. Test to see if sample x is consistent with an Exp(1) distribution using a QQ plot

plot(qexp(ppoints(x),1),sort(x))
abline(0,1)

Examine the help for the function qqnorm.

10
 9. Compare the two samples with a QQ plot

qqplot(x,y)
abline(0,1)

10. Compare the two samples with box plots

boxplot(x,y)

11. A simple alternative to display the two samples

plot(c(x,y),rep(0:1,c(length(x),length(y))),xlab="",ylab="")

12. Plot a histogram of x and a box plot of y, in the same figure.

par(mfrow=c(2,1))
hist(x)
boxplot(y)

13. Consider the Pima Indians data: a collection of variables observed on a particular group of native
American Indians who are either healthy or diabetic. This data includes measurements of tricep,
skinfold and blood glucose level.
First, load the mlbench package.

attach(PimaIndiansDiabetes)
plot(triceps,glucose,type="n")
plot(triceps[diabetes=="neg"],glucose[diabetes=="neg"],xlab="Tricep",ylab="glucose")
points(triceps[diabetes=="pos"],glucose[diabetes=="pos"],xlab="Tricep",ylab="glucose",col=2,pch=2)
legend(40,50,c("Diabetes","Healthy"),pch=1:2)

Problems
1. Examine the built in ChickWeight data (the help gives background about the data). The function
split will prove useful to do the following (as will a script)

(a) Construct a plot of weight against time for chick number 34.
(b) For chicks in diet group 4, display box plots for each time point.
(c) Compute the mean weight for chicks in group 4, for each time point. Plot this mean value
against time.
(d) Repeat the previous computation for group 2. Add the mean for group 2 to the existing plot.
(e) Add a legend and a title.
(f) Copy and paste the graph into Word.

Note that some of these steps will be easier once we have some more programming expertise.

11
 Introduction to R
Exercise 8

This sheet is concerned with control structures.

1. Write an R expression to determine if two sets, A and B, represented as integer vectors are disjoint.
If they are disjoint, display elements of set A otherwise display elements of set B. (Examine the
help for functions print and cat).

2. Write R codes that takes the coefficients of a quadratic equation, and outputs an appropriate
message for the cases of (i). two distinct roots (b2 − 4ac > 0) (ii) coincident roots (b2 = 4ac) or
(iii). complex roots (b2 < 4ac).

3. Let vector y be the logarithm of a random sample from a standard normal distribution, N (0, 1).
Use the ifelse function to replace missing values with the value 9999.

4. Let n be a large integer. Compute a vector x containing n random uniform deviates. Embed the
following code in the system.time function

for (i in 1:n) y[i] <- sin(x[i])

Do this a few times. Now, time the call

y <- sin(x)

Which is faster?

5. Compound interest can be computed using the formula

A = P × (1 + R/100)n

where P is the original money lent, A is what it amounts to in n years at R percent per year
interest.
Write R code to calculate the amount of money owed after n years, where n changes from 1 to
15 in yearly increments, if the money lent originally is 5000 pounds and the interest rate remains
constant throughout the period at 11.5%.

6. Write a loop structure to scan through an integer vector to determine the index of the maximum
value. The loop should terminate as soon as the index is obtained. (Don’t worry about ties).
Of course, this is not a good way to do this! Examine the help for the rank, sort and order
functions.

12