An introduction to R

• Basic types
• Basic operators
• Variables
• Libraries
• The pipe operator
• Coding style

More complex data types

• Vectors
• Factors
• Matrices
• Arrays
• Lists
• Data Frames

Vectors are ordered list of values.

• Vectors can be of any data type
  • numeric
  • character
  • logic
• All items in a vector have to be of the same type
• Vectors can be of any length

A vector variable can be defined using

• an identifier (e.g., a_vector)
• on the left of an assignment operator <-
• followed by the object to be linked to the identifier
• in this case, the result returned by the function c
• which creates a vector containing the element provided as input

a_vector <- c("Birmingham", "Derby", "Leicester",
  "Lincoln", "Nottingham", "Wolverhampton")
a_vector

## [1] "Birmingham"    "Derby"         "Leicester"     "Lincoln"      
## [5] "Nottingham"    "Wolverhampton"

• the operator :
• the function seq
• the function rep

4:7

## [1] 4 5 6 7

seq(1, 7, by = 0.5)

##  [1] 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0

seq(1, 10, length.out = 7)

## [1]  1.0  2.5  4.0  5.5  7.0  8.5 10.0

rep("Ciao", 4)

## [1] "Ciao" "Ciao" "Ciao" "Ciao"

Each element of a vector can be retrieved specifying the related index between square brackets, after the identifier of the vector. The first element of the vector has index 1.

a_vector[3]

## [1] "Leicester"

A vector of indexes can be used to retrieve more than one element.

a_vector[c(5, 3)]

## [1] "Nottingham" "Leicester"

Functions can be used on a vector variable directly

a_numeric_vector <- 1:5
a_numeric_vector + 10

## [1] 11 12 13 14 15

sqrt(a_numeric_vector)

## [1] 1.000000 1.414214 1.732051 2.000000 2.236068

a_numeric_vector >= 3

## [1] FALSE FALSE  TRUE  TRUE  TRUE

Overall expressions can be tested using the functions:

• any, TRUE if any of the elements satisfies the condition
• all, TRUE if all of the elements satisfy the condition

any(a_numeric_vector >= 3)

## [1] TRUE

all(a_numeric_vector >= 3)

## [1] FALSE

A factor is a data type similar to a vector. However, the values contained in a factor can only be selected from a set of levels.

houses_vector <- c("Bungalow", "Flat", "Flat",
  "Detached", "Flat", "Terrace", "Terrace")
houses_vector

## [1] "Bungalow" "Flat"     "Flat"     "Detached" "Flat"     "Terrace" 
## [7] "Terrace"

houses_factor <- factor(c("Bungalow", "Flat", "Flat",
  "Detached", "Flat", "Terrace", "Terrace"))
houses_factor

## [1] Bungalow Flat     Flat     Detached Flat     Terrace  Terrace 
## Levels: Bungalow Detached Flat Terrace

The function table can be used to obtain a tabulated count for each level.

houses_factor <- factor(c("Bungalow", "Flat", "Flat",
  "Detached", "Flat", "Terrace", "Terrace"))
houses_factor

## [1] Bungalow Flat     Flat     Detached Flat     Terrace  Terrace 
## Levels: Bungalow Detached Flat Terrace

table(houses_factor)

## houses_factor
## Bungalow Detached     Flat  Terrace 
##        1        1        3        2

A specific set of levels can be specified when creating a factor by providing a levels argument.

houses_factor_spec <- factor(
  c("People Carrier", "Flat", "Flat", "Hatchback",
      "Flat", "Terrace", "Terrace"),
  levels = c("Bungalow", "Flat", "Detached",
       "Semi", "Terrace"))

table(houses_factor_spec)

## houses_factor_spec
## Bungalow     Flat Detached     Semi  Terrace 
##        0        3        0        0        2

In statistics terminology, (unordered) factors are categorical (i.e., binary or nominal) variables. Levels are not ordered.

income_nominal <- factor(
  c("High", "High", "Low", "Low", "Low",
      "Medium", "Low", "Medium"),
  levels = c("Low", "Medium", "High"))
income_nominal > "Low"

## Warning in Ops.factor(income_nominal, "Low"): '>' not meaningful for
## factors

## [1] NA NA NA NA NA NA NA NA

In statistics terminology, ordered factors are ordinal variables. Levels are ordered.

income_ordered <- ordered(
  c("High", "High", "Low", "Low", "Low",
      "Medium", "Low", "Medium"),
  levels = c("Low", "Medium", "High"))
income_ordered > "Low"

## [1]  TRUE  TRUE FALSE FALSE FALSE  TRUE FALSE  TRUE

sort(income_ordered)

## [1] Low    Low    Low    Low    Medium Medium High   High  
## Levels: Low < Medium < High

Matrices are collections of numerics arranged in a two-dimensional rectangular layout

• the first argument is a vector of values
• the second specifies number of rows and columns
• R offers operators and functions for matrix algebra

a_matrix <- matrix(c(3, 5, 7, 4, 3, 1), c(3, 2))
a_matrix

##      [,1] [,2]
## [1,]    3    4
## [2,]    5    3
## [3,]    7    1

Variables of the type array are higher-dimensional matrices.

• the first argument is a vector containing the values
• the second argument is a vector specifying the depth of each dimension

a3dim_array <- array(1:24, dim=c(4, 3, 2))

a3dim_array

## , , 1
## 
##      [,1] [,2] [,3]
## [1,]    1    5    9
## [2,]    2    6   10
## [3,]    3    7   11
## [4,]    4    8   12
## 
## , , 2
## 
##      [,1] [,2] [,3]
## [1,]   13   17   21
## [2,]   14   18   22
## [3,]   15   19   23
## [4,]   16   20   24

Subsets of matrices (and arrays) can be selected as seen for vectors.

a_matrix[2, c(1, 2)]

## [1] 5 3

a3dim_array[c(1, 2), 2, 2]

## [1] 17 18

apply applies another function to each level of a set dimension of an array

apply(a3dim_array, 3, min) # apply on third dimension

## [1]  1 13

apply(a3dim_array, 1, min) # apply on first dimension

## [1] 1 2 3 4

apply(a3dim_array, 2, min) # apply on second dimension

## [1] 1 5 9

Variables of the type list can contain elements of different types (including vectors and matrices), whereas elements of vectors are all of the same type.

employee <- list("Stefano", 2015)
employee

## [[1]]
## [1] "Stefano"
## 
## [[2]]
## [1] 2015

employee[[1]] # Note the double square brackets for selection

## [1] "Stefano"

In named lists each element has a name, and elements can be selected to using their name after the symbol $.

employee <- list(name = "Stefano", start_year = 2015)
employee

## $name
## [1] "Stefano"
## 
## $start_year
## [1] 2015

employee$name

## [1] "Stefano"

With lapply take care that the function makes sense for any element in the list

various <- list(
  "Some text",
  matrix(c(6, 3, 1, 2), c(2, 2))
)
lapply(various, is.numeric)

## [[1]]
## [1] FALSE
## 
## [[2]]
## [1] TRUE

A data frame is equivalent to a named list where all elements are vectors of the same length.

employees <- data.frame(
  Name = c("Maria", "Pete", "Sarah"),
  Age = c(47, 34, 32),
  Role = c("Professor", "Researcher", "Researcher"))
employees

##    Name Age       Role
## 1 Maria  47  Professor
## 2  Pete  34 Researcher
## 3 Sarah  32 Researcher

Data frames are the most common way to represent tabular data in R. Matrices and lists can be converted to data frames.

Selection is similar to vectors and lists.

employees[1, ] # row selection

##    Name Age      Role
## 1 Maria  47 Professor

employees[, 1] # column selection, as for matrices

## [1] Maria Pete  Sarah
## Levels: Maria Pete Sarah

Selection is similar to vectors and lists.

employees$Name # column selection, as for named lists

## [1] Maria Pete  Sarah
## Levels: Maria Pete Sarah

employees$Name[1]

## [1] Maria
## Levels: Maria Pete Sarah

Values can be assigned to cells through filtering and <-

employees$Age[3] <- 33 
employees

##    Name Age       Role
## 1 Maria  47  Professor
## 2  Pete  34 Researcher
## 3 Sarah  33 Researcher

Operations can be performed on columns, and new columns created.

current_year <- as.integer(format(Sys.Date(), "%Y"))
employees$Year_of_birth <- current_year - employees$Age
employees

##    Name Age       Role Year_of_birth
## 1 Maria  47  Professor          1973
## 2  Pete  34 Researcher          1986
## 3 Sarah  33 Researcher          1987

A tibble is a modern reimagining of the data.frame within tidyverse

• they do less
  • don’t change variable names or types
  • don’t do partial matching
• complain more
  • e.g. when a variable does not exist

This forces you to confront problems earlier, typically leading to cleaner, more expressive code.

More complex data types

• Vectors
• Factors
• Matrices
• Arrays
• Lists
• Data Frames

In the practical session, we will see (surprise, surprise)

• Vectors
• Factors
• Matrices
• Arrays
• Lists
• Data Frames

Moving on towards data science

• Data selection
• Data filtering
• Data manipulation
• Join operations
• Table re-shaping