R Mini Course – Course developed by Mike McCann (2015) and Ben Weinstein (2014)

View Github Repository Open presentation in a new window

nlk124

See all presentation from nlk124

R Mini Course – Course developed by Mike McCann (2015) and Ben Weinstein (2014)

0 0

Rminicourse

On Github nlk124 / Rminicourse

R Mini Course

Course developed by Mike McCann (2015) and Ben Weinstein (2014)

Nicole Kinlock

January 21st and 22nd, 2016

Directory

Part 1: Basics of R

Back to the directory

Introduction to R

  • Statistical programming language and environment
  • Available as public domain software (i.e. not copyrighted)
  • Free!
  • Better than most commercial alternatives
  • Available on all platforms (Windows, OSX, Linux)
  • Not just for statistics, but also general purpose programming

Why use R?

  • Flexible
  • Transparent
  • Large user base
  • Valued skill (Remember, R is a programming language)

RStudio

  • Integrated development environment (IDE) for the R language
  • AKA use R with a nice, shiny interface
  • Also free!
  • Using R without RStudio feels kind of like using only Command (Windows) or Terminal (Mac) on your computer

Running R with RStudio

  • The screen is split four ways:
    • Workspace: (Top left) Where scripts are written & saved
    • Console: (Bottom left) Where commands are run
    • Environment, History: (Top right) Where objects are stored
    • Files, Plots, Packages, and Help: (Bottom right) View data, help

Running R with RStudio

  • Type directly into the console (best when you don't want to save the code) or type into the script - then run (CTRL + ENTER) (Windows) or (CMD + ENTER) (Mac)

  • A script is a plain text file with R commands in it. This will be where you save the code that you are writing - the file will end in the extension .R

R as a calculator

  • R has many arithmetic operators

    • + Addition
    • - Subtraction
    • * Multiplication
    • / Division
    • ^ Exponentiation
    • %% Modulus (finds remainder)
    • %/% Integer division (leaves off remainder)

  • R obeys the standard order of operations

R as a calculator

Examples

7 + 4

## [1] 11

3^2

## [1] 9

10 %% 7

## [1] 3

R is logical

  • R also has many logical operators
    • < Less than
    • <= Less than or equal to
    • > Greater than
    • >= Greater than or equal to
    • == Exactly equal to
    • != Not equal to
    • ! NOT
    • | OR
    • & AND

R is logical

Examples

7 == 4

## [1] FALSE

3 > 2

## [1] TRUE

7!=4

## [1] TRUE

Try It!

What is 17 multiplied by 365?

What is 13 cubed?

Is 9 to the fourth equal to the sum of 2000 and 187 multiplied by 3?

Creating Objects

  • An object is the fundamental unit in R. All expressions can be saved as an object.

  • To create an object from an expression we use the assignment operator (<-). The assignment operator assigns values on the right to objects on the left.

a <- (12 + 180) * 3
a

## [1] 576
  • The object a is now the output of the expression (12 + 180) * 3. Check your environment (upper right panel)

Assignment operator

  • Do not use a = 12 + 180 for assignment in R. It's best practice to use <- intead of =.

  • This may seem arbitrary, but it is helpful when you are reading someone else's code.

Google's R Style Guide

Advanced R: Style Guide

R Tip: Comment on your code

Use # signs to comment on your script. Anything to the right of a # is ignored. Good scripts (and homework) have comments before every major block of code. It's surprisingly hard to remember what you did when reviewing older code without comments, and it's particularly important when other people are reading your code. 

5 + 5 # This adds five and five 

## [1] 10

# 10 + 10 this does not add ten and ten

Expressions using objects

  • Objects can be combined into other, larger, and more complex objects.
a <- 8 * 10
b <- 2 * 10
d <- a * b
d

## [1] 1600

# This is equivalent to: 
d <- 8 * 10 * 2 * 10
d

## [1] 1600

Try It!

Create an object that is equal to your age. Create another object that is equal to the age of the person to your right. Find the difference between these objects.

Data structures: 1. Vectors

  • R has five common data structures. We will start with the simplest: vectors.
  • Vectors are one dimensional strings of numbers or characters. Vectors come in four types: numeric, integer, logical (TRUE or FALSE), and character. A vector is made using the combine function, c().

Data structures: 1. Vectors

Combining integers and numbers into vectors

a <- c(3, 4, 5)
a

## [1] 3 4 5

b <- c(3.24, 4.57, 5.03)
b

## [1] 3.24 4.57 5.03

Data structures: 1. Vectors

Combining characters into vectors

  • Characters in R need to be enclosed in quotation marks.
pets <- c("dog", "cat", "bird")
pets

## [1] "dog"  "cat"  "bird"

Data structures: 1. Vectors

Combining multiple objects into a single, new vector

# Make objects 
a <- sqrt(4 * 7)
b <- 6 * 5
g <- 9 * 2

# Combine
d <- c(a, b, g)
d

## [1]  5.291503 30.000000 18.000000

Data structures: 1. Vectors

Vectors of regular sequences

  • You can use a colon (:) to create a vector that includes all integers in between the numbers on either side of the colon
x <- 1:10
x

##  [1]  1  2  3  4  5  6  7  8  9 10
  • You can use seq(from = , to = , by = ) to create a vector with a set min and max (from, to) with a specified increment (by)
x <- seq(from = 1, to = 20, by = 2)
x

##  [1]  1  3  5  7  9 11 13 15 17 19

Try It!

Create a vector that includes all integers between 2 and 22. Name this vector x. Create a vector that includes all even integers between 2 and 22. Name this vector y. Create a vector that combines object x and object y. Name this vector z.

Data structures: 1. Vectors

Vector indexing: positional

  • You can access any element in the vector by putting its position in square brackets [ ] 
# Create a vector 
height <- c(76, 72, 74, 74, 78)
height

## [1] 76 72 74 74 78

height[1] # extract the 1st element in the vector 

## [1] 76

height[5] # extract the 5th element

## [1] 78

Data structures: 1. Vectors

Vector indexing: positional

  • You can also use vector indexing to return the same vector with certain elements missing using the - symbol
height <- c(76, 72, 74, 74, 78)

height[-1]

## [1] 72 74 74 78

Data structures: 1. Vectors

Vector indexing: named

  • You can assign names to each element of the vector, and then extract the element by indexing based on the name.
# Create a vector with named elements 
temp <- c(monday = 28.1, tuesday = 28.5, wednesday = 29.0, thursday = 30.1, friday = 30.2)
temp

##    monday   tuesday wednesday  thursday    friday 
##      28.1      28.5      29.0      30.1      30.2

temp["wednesday"]

## wednesday 
##        29

temp[3]

## wednesday 
##        29

Data structures: 1. Vectors

Vector indexing: logical

  • You can extract elements in a vector that meet specific criteria based on a logical expression.
y <- 5:50
y

##  [1]  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
## [24] 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

y[y <= 10]  # extract all elements less than or equal to 10

## [1]  5  6  7  8  9 10

y[y < 10 & y != 5]  # extract all elements less than 10 that are not equal to 5

## [1] 6 7 8 9

Try It!

What are the 9th and 12th positions of the vector seq(1, 27, 0.5)?

Bonus! Can you find those positions simultaneously?

Create the vector c(3:33) and name it a. Extract all elements of a that are greater than or equal to 17.

Functions

  • A function is a stored object that performs a task given some inputs (called arguments). R has many functions already available, but you can also write your own functions.

  • Try using the tab key while entering arguments in any function to discover a useful feature of RStudio.

  • Functions are used in the format: name_of_function(inputs)

  • The output of a function can be saved to an object: output <- name_of_function(inputs)

Functions

  • It's not necessary to explicitly name arguments, but it's helpful when learning the language.
seq(1, 10, 1)

##  [1]  1  2  3  4  5  6  7  8  9 10

seq(from = 1, to = 10, by = 1)

##  [1]  1  2  3  4  5  6  7  8  9 10

Functions

  • Use sum() to take the sum of all elements in a vector:
sum(c(3, 4, 5))

## [1] 12
  • Use mean() to take the mean of all elements in a vector:
mean(seq(5, 100, 5))

## [1] 52.5

Functions

  • Functions can act on an object
x <- seq(5, 100, 5)

# use the vector x as the input to the function
mean(x)

## [1] 52.5

R Tip: The help system

  • Help files provide important information on what the function does, how it works, and they provide examples at the very bottom.
help(mean)
  • You can also use ? before a function name to view the help file 
?mean  # Same as help(mean)
  • Use ?? to search for functions; e.g. search for any function whose help files contain the word "robust" 
??robust

Try It!

What is the median of 34, 16, 105, and 27?Remember: functions are often named intuitively.

What does the function range() do, and what is the sample example in the help file?

Is mean(4, 5) different than mean(c(4, 5))?

Packages

  • We will be exploring functions in much greater detail throughout this course. (Including writing your own functions!)

  • Functions are kept inside packages, some of which come pre-installed with R. Others must be downloaded.

Packages

  • There are many, many R packages - currently 7742!

  • Check the List of R Packages and search with your favorite keyword, e.g. ecology, paleo, dispersal, population, time series, phylogeny, community, Bayes

Installing Packages

  • Often you will need to install a package to access a certain library of functions.
# Install a new package
install.packages("picante")
  • Remember to surround the package name in quotation marks.

Loading Packages

  • Installing a package downloads it to your computer. You need to download packages before using them.

  • Additionally, you have to load packages before you use them every time you restart R. This lets R know what packages to load in, and not waste memory by loading all potential functions. 

# Two ways to load packages:
library(ggplot2)
require(ggplot2)
  • Note: no quotation marks needed

R Tip: Loading Packages

  • Good scripts (and homeworks) have a series of require() or library() statements at the top of the script.

  • Some say that you should preferentially use library(). The logic behind this is because require() "tries" to laod a package and library() actually loads a package. If you use require() at the top of the screen and you don't have the package installed, you'll get a failure message later on when you use functions in the package. However, if you use library() at the top of the screen and you don't have the package installed, you'll get a failure message right away.

Try It!

Search and find an interesting package. What is it? What is one function included in the package?

Install the package to your computer.

List of R Packages

The R User Community

Stack Overflow

R-Bloggers

R Mailing Lists

R for cats

Practice

Worksheets created by Mike McCann and hosted on his Rpubs site

Worksheet

Answers

Part 2: Probability & Distributions

Back to the directory

Review

What is the (sum of 1 through 10) multiplied by two?

What are the 12th and 45th positions of the vector seq(1, 43, 0.25)?

Find all of the elements of the vector seq(1, 10, 0.1) that are less than 2. What is the median value of these elements?

Distributions

  • It is often useful to generate a sample from a specific statisical distribution. To generate a sample of size 1000 from a standard normal distribution (mean 0 and standard deviation 1) we use the rnorm() function. 
norm <- rnorm(n = 1000, mean = 0, sd = 1)
head(norm)

## [1] -2.0902195660 -0.5281914723 -0.1530587059 -0.4630696300  0.3055771295
## [6]  0.0002270506

hist(norm, col = "palevioletred", main = "", xlab = "", ylab = "Frequency")

Try It!

Look up the rnorm() function in help screen. Locate the three arguments we used.

Draw 100 random numbers from a normal distribution with a mean of 3 and an sd of 2, assign it to an object a.

Find the mean of your sample. How close was it to the true mean?

What is the 13th element in your vector a?

Generating random samples from other distributions

  • R has many distributions in the base package, including all those commonly used in biological analysis.
  • Depending on the distribution, each function has its own set of parameter arguments.
  • For example, the rpois() function is the random number generator for the Poisson distribution and it has only the parameter lambda (\(\lambda\)).

BEE552 Students: Heather will pick up from here.

Try It!

Draw 100 random variates from a Poisson distribution with a lambda = 3, assign it to an object a.

Draw 1000 random variates from a Poisson distribution with a lambda = 3, assign it to an object b.

Find the means of both draws. What is the difference in means?

Other properties of distributions

  • For each distribution, there are four functions which will generate fundamental quantities of a distribution. Let's consider the normal distribution as an example.

    • rnorm() a random sample from a normal distribution.
    • dnorm() the density for a specific value for a normal distribution.
    • pnorm() the distribution function
    • qnorm() the quantile function

  • pnorm(), and qnorm() will be covered during Biometry. They are less commonly used.

Univariate plots

Barplots

  • Barplots are a basic type of univariate plot. The R function to create a barplot is barplot().
# Use the R dataset of the length of rivers in the United States (a univariate data set)
head(rivers)

## [1] 735 320 325 392 524 450

barplot(height = rivers[1:10], col = "paleturquoise", main = "Lengths of rivers in the United States", ylab = "Length (in miles)")

Univariate plots

Histograms

  • Histograms are also a common univariate plot. Histograms place data into "bins", and count the number (frequency) of data falling into each bin.

  • Bins are usually plotted as bars, with the x range on the x axis, and frequency on the y axis.

Univariate plots

Histograms

Histograms are an effective way of visualizing distributions

# Generate and visualize 100 random points from a standard normal distribution
sample <- rnorm(1000)
hist(sample, col = "palevioletred", main = "", xlab = "", ylab = "Frequency")

Try It!

Draw 10 random variates from a normal distribution and plot a histogram of the sample. Repeat for 100, then 1000 random variates. What do you notice about the histograms?

Explore at least one other distribution; look up ?distributions.

Take a random sample of the new distribution, plot the sample, and share with your neighbor.

Draw 1000 random variates from a normal distribution with a mean of 0 and a sd of 1. Look at ?hist. How do you specify the size of the bin range? Try making bins from -4 to 4, with intervals of 0.01, 0.1, and 1. Hint: Consider using seq() in the "breaks" argument within hist().

Sampling

  • In R, it's very easy to take a random sample of elements in any vector with the sample() function.

  • Take a random sample of 20 elements from a vector of integers between 1 and 100.

x <- 1:100
sample(x, size = 20) # Sample without replacement

##  [1] 96 56 10 30  4 99 93 60 49 36 85 65 28 88  2 32 48 64 94 76

sample(x, size = 20, replace = TRUE) # Sample with replacement

##  [1] 24 59 59 97  7 78 81 10 68 31 42 50 46 32 87 42 84 57 46  5

Try It!

Sample 14 elements from a vector of integers between 100 and 200 without replacement.

Sample 5 letters (using the pre-installed vector letters) from the alphabet with replacement.

Sample 0 or 1 twelve times, with and without replacement. What happens?

Bivariate plots

Scatterplots

  • Scatterplots are useful for showing the relationship between two variables 
sample1 <- rnorm(n = 40, mean = 4, sd = 2)
sample2 <- rnorm(n = 40, mean = 2, sd = 1)
plot(formula = sample2 ~ sample1, pch = 20)

Bivariate plots

Scatterplots

  • Instead of writing the relationship as a formula, i.e. y ~ x, you can write plot(x, y).
plot(x = sample1, y = sample2, pch = 20)

Bivariate plots

Scatterplots

  • You can add straight lines to a plot with abline(). a specifies the intercept and b the slope.
plot(sample1, sample2)
abline(a = -0.1, b = 1, col="royalblue4")

Bivariate plots

Scatterplots

  • Lines can also be model fits

  • lm() fits a linear relationship between x and y.

plot(sample1, sample2)
abline(lm(sample2 ~ sample1), col="royalblue4")

Lots of useful links for plotting

R Base Graphics: An Idiot's Guide

R Base Graphics Plotting

Quick-R: Basic Graphs

Producing Simple Graphs with R

R Colors

More on plotting later (i.e. making plots that are visually appealing!)

Practice

Worksheet

Answers

Part 3: Data Manipulation

Back to the directory

Download the file for this section

seedlings.csv

Data structures: 2. Data frames

  • Most data you've worked with probably isn't in vector form. More likely, it includes multiple rows and columns. Thus, a data frame is a very important data type in R and what you'll be working with much of the time. 
head(trees)  # Pre-loaded dataset in R - measurements of black cherry trees

##   Girth Height Volume
## 1   8.3     70   10.3
## 2   8.6     65   10.3
## 3   8.8     63   10.2
## 4  10.5     72   16.4
## 5  10.7     81   18.8
## 6  10.8     83   19.7

Data structures: 2. Data frames

Data frames are usually read in from a file, but R comes with many practice datasets. We will use the iris dataset, famously used by R.A. Fisher in 1936

head(iris)

##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
## 1          5.1         3.5          1.4         0.2  setosa
## 2          4.9         3.0          1.4         0.2  setosa
## 3          4.7         3.2          1.3         0.2  setosa
## 4          4.6         3.1          1.5         0.2  setosa
## 5          5.0         3.6          1.4         0.2  setosa
## 6          5.4         3.9          1.7         0.4  setosa

Data structures: 2. Data frames

Useful functions for data frames

  • head() - see first 5 rows
  • tail() - see last 5 rows
  • dim() - dimensions (# rows, # columns)
  • nrow() - number of rows
  • ncol() - number of columns
  • str() - structure of any object
  • class() - class of any object
  • rownames() - row names
  • colnames() - column names

Try It!

How many rows does the iris data frame have?

How many columns? What are the column names?

Using the str() function, how many species are in the data?

What classes are each of the columns?

Data structures: 2. Data frames

Data frame syntax and subsetting

  • R has many powerful subset operators that will allow you to easily perform complex operations on any kind of dataset (much easier than with Excel!)
  • Think of data frames as a series of vectors grouped together into a table
head(iris)

##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
## 1          5.1         3.5          1.4         0.2  setosa
## 2          4.9         3.0          1.4         0.2  setosa
## 3          4.7         3.2          1.3         0.2  setosa
## 4          4.6         3.1          1.5         0.2  setosa
## 5          5.0         3.6          1.4         0.2  setosa
## 6          5.4         3.9          1.7         0.4  setosa

Data structures: 2. Data frames

Data frame indexing: positional

  • Index individual elements of a data frame with nameofdf[row#, column#] 
iris[1, 1]  # first row, first column

## [1] 5.1

iris[3, 3]  # third row, third column

## [1] 1.3

Data structures: 2. Data frames

Data frame indexing: positional

  • Index entire rows or columns
iris[2, ]  # subset the entire second row 

##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
## 2          4.9           3          1.4         0.2  setosa

iris[, 2]  # subset the entire second column 

##   [1] 3.5 3.0 3.2 3.1 3.6 3.9 3.4 3.4 2.9 3.1 3.7 3.4 3.0 3.0 4.0 4.4 3.9
##  [18] 3.5 3.8 3.8 3.4 3.7 3.6 3.3 3.4 3.0 3.4 3.5 3.4 3.2 3.1 3.4 4.1 4.2
##  [35] 3.1 3.2 3.5 3.6 3.0 3.4 3.5 2.3 3.2 3.5 3.8 3.0 3.8 3.2 3.7 3.3 3.2
##  [52] 3.2 3.1 2.3 2.8 2.8 3.3 2.4 2.9 2.7 2.0 3.0 2.2 2.9 2.9 3.1 3.0 2.7
##  [69] 2.2 2.5 3.2 2.8 2.5 2.8 2.9 3.0 2.8 3.0 2.9 2.6 2.4 2.4 2.7 2.7 3.0
##  [86] 3.4 3.1 2.3 3.0 2.5 2.6 3.0 2.6 2.3 2.7 3.0 2.9 2.9 2.5 2.8 3.3 2.7
## [103] 3.0 2.9 3.0 3.0 2.5 2.9 2.5 3.6 3.2 2.7 3.0 2.5 2.8 3.2 3.0 3.8 2.6
## [120] 2.2 3.2 2.8 2.8 2.7 3.3 3.2 2.8 3.0 2.8 3.0 2.8 3.8 2.8 2.8 2.6 3.0
## [137] 3.4 3.1 3.0 3.1 3.1 3.1 2.7 3.2 3.3 3.0 2.5 3.0 3.4 3.0

Data structures: 2. Data frames

Data frame indexing: positional

  • Data frames can be indexed for both rows and columns

  • Get the 5th, 7th, and 9th rows for the first two columns.

iris[c(5, 7, 9), 1:2]

##   Sepal.Length Sepal.Width
## 5          5.0         3.6
## 7          4.6         3.4
## 9          4.4         2.9

Data structures: 2. Data frames

Data frame indexing: named

  • Columns can also be called by name using brackets
iris[, "Sepal.Length"]
  • Or, use the dollar sign $ to call a column by name
iris$Sepal.Length

Try It!

What is the 9th entry of the Sepal.Width column? Call it x.

Subset the 17th row of the data frame iris.

Return an object with the 1st, 4th and 7th rows of the iris dataframe.

Use the seq() function to subset all odd rows in the iris dataset.

What happens when you use negative numbers to index the iris dataframe? Hint: Use dim() on the original and final objects.

Data structures: 2. Data frames

Data frame indexing: logical

  • An extremely helpful tool is to subset your data using logic rather than an index.
petal <- iris$Petal.Width  # Subset of the column petal width

hist(petal, col = "darkseagreen1", main = "Petal width of Iris", xlab = "", ylab = "Frequency") # Make a histogram of petal width frequency

Data structures: 2. Data frames

Data frame indexing: logical

logi <- petal > 1 # Which petal widths are greater than 1?
head(logi)

## [1] FALSE FALSE FALSE FALSE FALSE FALSE

iris.subset <- iris[logi, ] # Subset of iris only including individuals
# where petal width is greater than 1

head(iris.subset)

##    Sepal.Length Sepal.Width Petal.Length Petal.Width    Species
## 51          7.0         3.2          4.7         1.4 versicolor
## 52          6.4         3.2          4.5         1.5 versicolor
## 53          6.9         3.1          4.9         1.5 versicolor
## 54          5.5         2.3          4.0         1.3 versicolor
## 55          6.5         2.8          4.6         1.5 versicolor
## 56          5.7         2.8          4.5         1.3 versicolor

This is the same as:

iris[iris$Petal.Width > 1, ]

Try It!

Why is iris[iris > 3, ] a nonsensical command?

What about iris[iris$Sepal.Length > 3]?

Create a histogram of petal lengths for the entire data.

Subset the data for petal lengths greater than two.

Create a histogram of your new data.

Data structures: 2. Data frames

Data frame indexing: logical

  • For all types of data, if we want to subset a specific value/character/factor we use == 
iris.subset <- iris[iris$Petal.Length == 4,]
head(iris.subset)

##    Sepal.Length Sepal.Width Petal.Length Petal.Width    Species
## 54          5.5         2.3            4         1.3 versicolor
## 63          6.0         2.2            4         1.0 versicolor
## 72          6.1         2.8            4         1.3 versicolor
## 90          5.5         2.5            4         1.3 versicolor
## 93          5.8         2.6            4         1.2 versicolor

Data structures: 2. Data frames

Data frame indexing: logical

  • Subset only records from the species I. versicolor 
versicolor_only <- iris[iris$Species == "versicolor", ]
head(versicolor_only)

##    Sepal.Length Sepal.Width Petal.Length Petal.Width    Species
## 51          7.0         3.2          4.7         1.4 versicolor
## 52          6.4         3.2          4.5         1.5 versicolor
## 53          6.9         3.1          4.9         1.5 versicolor
## 54          5.5         2.3          4.0         1.3 versicolor
## 55          6.5         2.8          4.6         1.5 versicolor
## 56          5.7         2.8          4.5         1.3 versicolor

Data structures: 2. Data frames

Data frame indexing: logical

  • It's also possible to combine logical statments:
    • Use & (AND) if you want both statements to be true.
    • Use | (OR) if you want either statement to be true.
versicolor.4 <- iris[iris$Petal.Length > 4 & iris$Species == "versicolor", ]  # subset of observations or only I. versicolor where petal length is greater than 4
head(versicolor.4)

##    Sepal.Length Sepal.Width Petal.Length Petal.Width    Species
## 51          7.0         3.2          4.7         1.4 versicolor
## 52          6.4         3.2          4.5         1.5 versicolor
## 53          6.9         3.1          4.9         1.5 versicolor
## 55          6.5         2.8          4.6         1.5 versicolor
## 56          5.7         2.8          4.5         1.3 versicolor
## 57          6.3         3.3          4.7         1.6 versicolor

Try It!

Explain in words each of the following logical statements

iris[1:4, ]

iris[c(1:15), c(1, 3)]

iris[iris$Species == "setosa", "Petal.Width"]

What happens when you add a ! before a logical statment? Hint: Compare iris[iris$Species == "setosa", ] and iris[!iris$Species == "setosa", ].

Data structures: 2. Data frames

Build a data frame from scratch

  • Use the data.frame function. 
df <- data.frame(x = 1:5, y = 6:2)
df

##   x y
## 1 1 6
## 2 2 5
## 3 3 4
## 4 4 3
## 5 5 2

Data structures: 2. Data frames

Add columns to data frames

  • Use the data.frame function (again).
df <- data.frame(df, z = 41:45)
df

##   x y  z
## 1 1 6 41
## 2 2 5 42
## 3 3 4 43
## 4 4 3 44
## 5 5 2 45
  • Or, use the assignment operator <- 
df$z <- 41:45
df

##   x y  z
## 1 1 6 41
## 2 2 5 42
## 3 3 4 43
## 4 4 3 44
## 5 5 2 45

Data structures: 2. Data frames

Remove columns from data frames

  • Use the - symbol and the positional index
df.subset <- df[, -3]
df.subset

##   x y
## 1 1 6
## 2 2 5
## 3 3 4
## 4 4 3
## 5 5 2
  • Or, assign NULL to the column
df$z <- NULL
df

##   x y
## 1 1 6
## 2 2 5
## 3 3 4
## 4 4 3
## 5 5 2

Data structures: 3. Matrices

  • Similar to data frames. Matrices are two-dimensional and only consist of numbers.

  • Why do you need them? Some functions require a matrix as an input. 

a <- matrix(1:9, ncol = 3, nrow = 3)
a

##      [,1] [,2] [,3]
## [1,]    1    4    7
## [2,]    2    5    8
## [3,]    3    6    9

Data structures: 3. Matrices

  • Indexing is the same as with data frames
a <- matrix(1:9, ncol = 3, nrow = 3)
a

##      [,1] [,2] [,3]
## [1,]    1    4    7
## [2,]    2    5    8
## [3,]    3    6    9

a[2, 2]

## [1] 5

a[2, ]

## [1] 2 5 8

Importing your own data

  • R is not a spreadsheet program, so it's not great for direct data entry. It's best to start with spreadsheets for data entry and storage, and to import spreadhseets into R for data visualization and analysis.

  • .csv (comma separated values) files are often the preferred format to import into R.

  • Before we do that, we will need to look at concept of your working directory.

Working directory

  • Find out what your current working directory is
getwd()

## [1] "/Users/nicolekinlock/Documents/Biometry TA/R Mini Course/slides/Rminicourse"
  • This is the folder on your computer where R will look to open or write files.

Working directory

  • Set your working directory 
setwd("/Users/nicolekinlock/Documents/Biometry TA/R Mini Course/slides/")

Importing your own data

  • Store the downloaded seedling data in your environment
seedlings <- read.csv("/Users/nicolekinlock/Documents/Biometry TA/R Mini Course/data/seedlings.csv")
head(seedlings)

R Tip: Variable names

  • One of the largest sources of frustration with R can be importing data. Variable names often cause problems.
    • Do not have spaces in variable names
    • Use lower case letters
    • Abbreviate when appropriate
Average Height # BAD - this won't work
Average.Height # OK - this will work but isn't best practice (See style guide) 
average.height # BETTER - this will work, but will be slow to type repeatedly
avg.height     # GOOD!

Try It!

Import the data frame seedlings.csv and store it. What is the mean height for all seedlings? What about the mean height for all seedlings except Rhus typhina?

Exporting data frames

  • After you've made changes (done subsetting, etc.), you may want to export the altered data frame.
# Write the file 
write.csv(iris, file = "iris.csv", row.names = FALSE)

# Check in your working directory for the new file
list.files()

Helpful links

Data structures

Practice

Worksheet

Answers

Part 4: Looping and vectorization

Back to the directory

For loops

  • Loops are an important programming tool. The first loop we will learn is a for loop.

  • For loops run for a certain number of steps (iterations) that you define, during which any statements in the loop are executed.

-The basic syntax is:

for (i in 1:number_of_iterations) { execute these statements }

For loops

Why use a for loop?

We have a repeated process with indentical formatting, but different values.

To avoid laborious typing into R

For loops

Our first loop

for (i in 1:5) {

  • i starts at 1. R will execute some statements;
  • i is increased to i = 2 and statements are executed again;
  • i is increased to i = 3 and statements are executed again;
  • and so on, until i = 5, at which point the loop executes the set of statements for the last time.

}

For loops

Our first loop

for (i in 1:5){
    print(i)
}

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

For loops

Syntax

Be sure you distinguish between:

  • curly braces { }
  • parentheses ( )
  • square brackets [ ].

For loops

Syntax

Brackets [ ] are used to access elements of vectors, matrices, and dataframes. 

x <- 1:10

x[6]

## [1] 6

x[6:9]

## [1] 6 7 8 9

For loops

Syntax

Parentheses ( ) are used to specify arguments to functions. 

x <- 1:10

sum(x)

## [1] 55

mean(x)

## [1] 5.5

For loops

Syntax

Finally, use curly braces { } to enclose all of the statements to be executed in a loop.

for (i in 1:3) {
  print(i)
}

## [1] 1
## [1] 2
## [1] 3

For loops

Using a for loop

You can perform operations on i.

for (i in 1:4){
    print(i^2)
}

## [1] 1
## [1] 4
## [1] 9
## [1] 16

For loops

Using a for loop

Assignments can occur in a loop. 

x <- 2

for (i in 1:4){
    x <- x^2
    print(x)
}

## [1] 4
## [1] 16
## [1] 256
## [1] 65536

Note: i is not directly called in the equation.

The operation x <- x^2 will be done four times. x changes with each iteration because of the re-assignment in the loop.

Try It!

Create a for loop that prints numbers 1 to 100.

Create a for loop that prints numbers 100 to 1.

Create a for loop that adds 1 to numbers 1:5.

Create a for loop that divides all even numbers from 0 to 20 by 10 (consider using seq()).

Bonus! What would be the final value here (witout trying it)? Why?

dogs <- 10
for (i in 1:5){
    dogs <- dogs + 1
}

For loops

For loops and vectors

In the above examples, we used i directly in mathematical operations. It is more common to loop over elements of a vector to accomplish some particular task.

nameVector <- c("Harry", "Hermione", "Ron")

for (i in 1:length(nameVector)){
    print(paste("Hi,", nameVector[i], sep=" "))
}

## [1] "Hi, Harry"
## [1] "Hi, Hermione"
## [1] "Hi, Ron"

For loops

What did that really do?

  • Consider the loop in pieces
length(nameVector) # The # of positions in nameVector

## [1] 3

nameVector[1] # The 1st position in nameVector

## [1] "Harry"

# Combine text and index of a vector
paste("Hi,", nameVector[1], sep=" ") 

## [1] "Hi, Harry"

For loops

What did that really do?

Loops are their own little environment, so use print() to view them on your console so you can see the output of each iteration. 

nameVector <- c("Harry", "Hermione", "Ron")

for (i in 1:length(nameVector)){
    print(paste("Hi,", nameVector[i], sep=" "))
}

## [1] "Hi, Harry"
## [1] "Hi, Hermione"
## [1] "Hi, Ron"

Without print() or an assignment <- results are not returned.

nameVector <- c("Harry", "Hermione", "Ron")

for (i in 1:length(nameVector)){
    paste("Hi,", nameVector[i], sep=" ")
}

Try It!

Create a vector of names of people in your row, write them a nice message using a loop.

Explain why the following code is wrong: 

for (x in 1:10) {
  print(sum(i))
}

Data structures: 4. Lists

  • Lists are another of the 5 basic data structures in R. Lists, like vectors, are 1 dimensional.

  • The elements of a vector must all b the same type. In this example, 1 and 2 are converted to characters because "blue" is included.

x <- c(1, 2, "blue")
x

## [1] "1"    "2"    "blue"

typeof(x)

## [1] "character"

Data structures: 4. Lists

  • In contrast, the elements of a list can be any type (numeric, integer, logical, character), or other lists.
a <- list(numeric = seq(0, 0.25, 0.01), integer = 10:20, logical = c(TRUE, FALSE, FALSE), character = c("open", "closed", "closed", "open"))
a

## $numeric
##  [1] 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13
## [15] 0.14 0.15 0.16 0.17 0.18 0.19 0.20 0.21 0.22 0.23 0.24 0.25
## 
## $integer
##  [1] 10 11 12 13 14 15 16 17 18 19 20
## 
## $logical
## [1]  TRUE FALSE FALSE
## 
## $character
## [1] "open"   "closed" "closed" "open"

Data structures: 4. Lists

Constructing lists and list syntax

  • You can construct a list with list() (instead of c() for vectors) 
x <- c(1, 2, 3, 4, 5)
y <- c(3, 6, 8, 10)

v <- c(x, y)  # Create a vector
str(v)

##  num [1:9] 1 2 3 4 5 3 6 8 10

l <- list(x, y)  # Create a list
str(l)

## List of 2
##  $ : num [1:5] 1 2 3 4 5
##  $ : num [1:4] 3 6 8 10

v[2]  # Index a vector 

## [1] 2

l[[2]]  # Index a list 

## [1]  3  6  8 10

Data structures: 4. Lists

List indexing

  • Using the single bracket indexing operator [] will return a list

  • Using a double bracket indexing operator [[]] will return contents within a list (integers, numbers, characters)

b <- list(1:4, c(3.25, 6.75, 8.175, 4.5), c("apples", "bananas", "oranges", "pears"))
b[2]  # returns a list

## [[1]]
## [1] 3.250 6.750 8.175 4.500

str(b[2])

## List of 1
##  $ : num [1:4] 3.25 6.75 8.18 4.5

b[[2]]  # returns numbers from list

## [1] 3.250 6.750 8.175 4.500

str(b[[2]])

##  num [1:4] 3.25 6.75 8.18 4.5

Data structures: 4. Lists

List indexing

b[1] * b[2]  # can't do arithmatic on lists

## Error in b[1] * b[2]: non-numeric argument to binary operator

b[[1]] * b[[2]]  # can do arithmatic on numbers

## [1]  3.250 13.500 24.525 18.000
  • In order to reference a single list element directly, we have to use the double square bracket [[ ]] operator and the single bracket [] operator. 
b[[2]] # 2nd element of list 

## [1] 3.250 6.750 8.175 4.500

b[[2]][2] # 2nd element of 2nd element of list 

## [1] 6.75

How do we save loop outputs?

  • Instead of printing to the screen, we usually want to create a new object with the outputs on the loop.
  • To do this, you need to create a blank vector or list before the loop
outputs <- c()  # Create a blank vector

x <- rnorm(n = 10, mean = 1, sd = 0.5)  # Vector that we will use in the loop

for (i in 1:length(x)) {
  outputs[i] <- x[i] * 10
}
outputs

##  [1]  6.029596 13.761718 10.922449 14.737241  6.798512  6.913397  1.505247
##  [8]  8.059382  7.972717  7.948766

Try It!

First, create a vector, y, with 10 draws from a standard normal distribution (mean = 0, sd = 1).

Compute y * 2 for 10 iterations (the length of vector y) using a for loop. Place the output in a vector.

Compute y * 2 for 10 iterations using a for loop. Place the output in a list.

Find the the 47th element in the output vector from question 1 and the output list from question 2.

What do each of these loops do? 

output <- list()
for(x in 1:10) {
  output[1] <- sum(x + x^2)
}

output2 <- list()
for(x in 1:11) {
  output2[x + 1] <- sum(x + x^2)
}

Flow Statements

if statements

if (3 > 2) {
  print("Yes")
}

## [1] "Yes"

Flow Statements

if statements

  • Often, we want for loops to be able to do different things under different conditions. We want the for loop to account for variables, options, and logical statements.

  • Let's use an if statement:

x <- 1:5

for (i in 1:length(x)) {
  if(x[i] > 3) {
    print(paste(x[i],"is greater than 3"))
  }
  if(x[i] <= 3) {
    print(paste(x[i],"is less than or equal to 3"))
  }
}

## [1] "1 is less than or equal to 3"
## [1] "2 is less than or equal to 3"
## [1] "3 is less than or equal to 3"
## [1] "4 is greater than 3"
## [1] "5 is greater than 3"

Flow Statements

Controlling flow: break statements

  • Sometimes we want the loop to end if certain logical criteria are met. We can end a loop running based on an if and break statement
for (i in 1:length(x)) {
  if(x[i] > 3){
    break
    }
  if(x[i] <= 3) {
    print(paste(x[i],"is less than or equal to 3"))
    }
}

## [1] "1 is less than or equal to 3"
## [1] "2 is less than or equal to 3"
## [1] "3 is less than or equal to 3"

Flow Statements

Controlling flow: next statements

  • Sometimes we don't want to break the statement, just skip a troublesome object or R that we know will cause an error.

  • We can continue within a loop based on an if and next statement. Here we want to skip 4. 

for (i in 1:length(x)) {
  if(x[i] == 4) {
    next
    }
  if(x[i] > 3) {
    print(paste(x[i],"is greater than 3"))
    }
  if(x[i] <= 3) {
    print(paste(x[i],"is less than or equal to 3"))
    }
  }

## [1] "1 is less than or equal to 3"
## [1] "2 is less than or equal to 3"
## [1] "3 is less than or equal to 3"
## [1] "5 is greater than 3"

Try It!

First, create a vector, x, with all integers between 1 and 100.

Create a for loop that computes x[i] * 2 for 100 iterations (the length of vector x). Place the output in a vector. However, calculate x[i] * 3 when x[i] = 32.

Create a for loop that computes x[i] * 2 for 100 iterations. Place the output in a list. However, break the loop after 51 iterations.

Create a for loop that computes x[i] * 2 for 100 iterations. Place the output in a vector. However, skip x = 71, 74.

Alternatives to for loops: the apply family

The apply family of functions allows you to process whole rows, columns, or lists. This is called vectorization. This can replace for loops (and is often faster).

apply()

  • apply(matrix_name, margin = c(1,2), function = function_name)
  • margin = 1 apply function to rows of matrix; margin = 2 apply to columns of matrix; margin = c(1,2) apply to rows and columns of matrix
  • apply(M, 2, fun) = apply fun to the columns of M

apply()

  • Use apply() to take the median of the rows/columns of matrix M.
M <- matrix(rnorm(9), ncol = 3, nrow = 3)
M

##            [,1]       [,2]        [,3]
## [1,]  0.1494313  0.8779963  0.25422087
## [2,] -0.8206744  1.4368966 -1.01193249
## [3,] -1.8166418 -0.4964814  0.06765827

apply(M, 1, median) # median of rows 

## [1]  0.2542209 -0.8206744 -0.4964814

apply(M, 2, median) # median of columns 

## [1] -0.82067442  0.87799633  0.06765827

apply()

  • The same analysis, but using a for loop
med.M <- c()

for (i in 1:3) {
    med.M[i] <- median(M[, i])
}

med.M

## [1] -0.82067442  0.87799633  0.06765827

The apply family

  • Other versions of apply() for other data types:

    • lapply()
    • tapply()
    • sapply()
    • vapply()
    • mapply()

  • Read more here and here.

The apply family

  • Other versions of apply() are used for other data types. For example lapply() uses lists as input.
mylist <- list(1:3, c(1.4, 4.6, 4.2, 2.1), rnorm(4)) # make a list 
mylist
sum(mylist)  # returns an error 

lapply(mylist, sum)  # applies function to each element in the list

Useful links

Subsetting tutorial

Loops in R

Apply family in R

Practice

Worksheet

Answers

Part 5: Functions

Back to the directory

Functions in R: Base

  • Functions contain sets of instructions that we want to carry out repeatedly

  • We have already seen many of the basic functions that come pre-installed with R.

sum(seq(1, 100, 1))
abs(-100 + 50)
dim(iris)
str(iris)
colnames(iris)

Functions in R: Packages

  • We have also seen functions that are loaded from packages. 
# This will not run. tree is not loaded 
tree(formula = Species ~ . -Species, data = iris) 

install.packages("tree") # Install package 
library(tree) # Load package 

# Now it should work 
tree(formula = Species ~ . -Species, data = iris)

Functions in R: Make your own!

It is also possible to define your own functions. This is especially important if you are going to write the same lines of code over and over again.

  • R functions are objects just like anything else.
  • By default, R function arguments are lazy, i.e. they're only evaluated if they're actually used.
  • Every call on a R object is almost always a function call.

Basic components of a function

  • The body(), the code inside the function.
  • The formals(), the "formal" argument list, which controls how you can call the function.
  • The environment(), which determines how variables referred to inside the function are found.
  • args() to list arguments.

Writing functions

function_name <- function(arguments) {body}

# define a function, f
f <- function(x, y){
  x + y
}

# call function f
f(x = 1, y = 3) 

## [1] 4

Functions and the environment

  • Variables defined inside functions exist in a different environment than the global environment, i.e. they don't exist outside of the function.

  • However, if a variable is not defined inside a function, the function will look one level above.

  • If you run this function, you'll see that y does not pop up in your environment panel.

x <- 2  # variable defined outside the function

g <- function() {
    y <- 1 # variable defined inside the function 
    c(x, y) 
}

g()

## [1] 2 1

Function basics

  • Unless you use the function return() or write the output at the end of the function, you won't be able to access the output
f1 <- function(a, b) {
    x <- a + b 
    y <- (a + b)^2
    z <- a/b
}

f1(1, 2)  # doesn't return anything

f2 <- function(a, b) {
    x <- a + b 
    y <- (a + b)^2
    z <- a/b
    c(x, y, z)  # same result as writing return(c(x, y, z))
}

f2(1, 2)  # returns x, y, and z

## [1] 3.0 9.0 0.5

Try It!

Create a function that takes in two arguments, x and y, and computes x * 2 * y.

Create a function that takes in three arguments, and makes a vector from the result.

Create a function that counts the number of matching elements in two separate vectors. Hint: use %in% to create a logical statement.

Function basics

  • Arguments can be passed to a function as a vector or a list
params <- c(5, 25)
params

## [1]  5 25

f3 <- function(p){
    alpha <- p[1]
    beta <- p[2]
    alpha * beta
}

f3(params)

## [1] 125

Functions with pre-defined values

subtract <- function(a = 5, b = 2){
    return(a - b)
}

subtract()

## [1] 3

subtract(5, 6)

## [1] -1

Try It!

Write a function that takes a vector as an argument and multiplies the sum of the vector by 10. Return a logical statement based on whether the sum is under 1000.

Write a function that calculates the mean of every column in a dataframe. Code the function so that it does not evaluate the column mean if the column elements are not numbers, using class(x) != "numeric". Try your function on the iris dataset.

Try It!

f4 <- function(x) {
  for (i in 1:ncol(x)) {
    if(class(x[, i]) != "numeric") {
      next
      }
    if(class(x[, i]) == "numeric") {
      print(mean(x[, i]))
    }
  }
}
f4(iris)

## [1] 5.843333
## [1] 3.057333
## [1] 3.758
## [1] 1.199333

Helpful links

Tutorial on using functions in R

Advanced R - Functions

Practice

Worksheet

Answers

Part 6: Data visualization using ggplot2

Back to the directory

Review

ggplot2 is not a base package, so we need to install it.

How would you install the package ggplot2?

What is the next step before you can use a function from ggplot2?

Install and load ggplot2 

install.packages("ggplot2")
library(ggplot2)

Why ggplot2?

  • The way you write code in ggplot2 makes a lot more sense compared to base graphics
  • Easier to create visually appealing figures
  • Useful for exploring trends in your data

Why ggplot2?

plot(x = iris$Species, y = iris$Sepal.Length)


library(ggplot2)
ggplot(data=iris, aes(x = Species, y = Sepal.Length)) + geom_boxplot() + theme_bw()

Why gg?

  • gg stands for "grammar of graphics"

  • Uses a set of terms that defines the basic components of (every) plot

  • Produce figures using coherent, consistent syntax (very similar code for very different figures)

    • However, this syntax does vary quite a bit from typical R syntax!
    • If you are a little confused by the syntax at first, that probably just means that you are getting more comfortable with the R language

The grammar

  • A basic ggplot2 plot consists of:
    • data: Must be a data frame
    • aes(thetics): How your data are represented visually (x, y, color, size, shape, etc.)
    • geom(etry): What type of geometric objects you want to display (points, lines, polygons, etc.)

A basic plot

library(ggplot2)
ggplot(data = iris, aes(x = Sepal.Length, y = Sepal.Width)) + geom_point()

Plots can be assembled in pieces

  • Useful for exploring data in different ways

  • I'll be using variables from the state.x77 and state.region datasets (US state data from 1977)

state <- data.frame(state.x77, state.region, state.abb)
stateplot <- ggplot(data = state, aes(x = Population, y = Area))
stateplot + geom_point()

Aesthetics

Increase the size of points

ggplot(data = state, aes(x = Population, y = Area)) + geom_point(size = 3)

Aesthetics

  • Differentiate state region by color 
ggplot(data = state, aes(x = Illiteracy, y = Murder, color = state.region)) + geom_point(size = 2)

Aesthetics

  • Differentiate state region by color and shape 
ggplot(data = state, aes(x = Life.Exp, y = HS.Grad, color = state.region, shape = state.region)) + geom_point(size = 2)

Try It!

  • Take a sample of the diamonds dataset
d2 <- diamonds[sample(1:nrow(diamonds), 1000), ]
  • Then generate this plot:

Other geometric objects

Type geom_ and hit tab to see them all. Then, use ?geom_nameofgeom to see the help screen.

Other geometric objects

Boxplot

ggplot(iris, aes(x = Species,y = Sepal.Length)) + geom_boxplot()

Other geometric objects

Histograms

  • Change some of the aesthetics - binwidth, fill 
ggplot(faithful, aes(x = waiting)) + geom_histogram(binwidth = 8, color = "black", fill = "paleturquoise")

Other geometric objects

Bar plots

ggplot(iris, aes(x = Species, y = Sepal.Length)) + geom_bar(stat = "identity")

Other geometric objects

Line plots

ggplot(mtcars, aes(x = wt, y = mpg, color = as.factor(cyl))) + geom_line()

Other geometric objects

Density plots

ggplot(faithful, aes(waiting)) + geom_density(fill = "thistle")

Try It!

Look up geom_histogram. What does it do?

Make a histogram of Sepal.Length from the iris data set. What did it do with the different species?

Facets

Plots can also have facets, which divide a plot into subplots based on some discrete variable (here, species). 

ggplot(iris, aes(Sepal.Length)) + geom_histogram() + facet_grid(Species ~ .)

Facets

Change to facet_grid(. ~ Species) and get one row, three columns.

ggplot(iris, aes(Sepal.Length)) + geom_histogram() + facet_grid(. ~ Species)

Stats

Type stat_ and hit tab to see them all. Then, use ?stat_nameofstat to see the help screen.

Stats

Use stat_smooth to add a linear fit

ggplot(iris, aes(x = Sepal.Length, y = Sepal.Width, color = Species)) + geom_point() + stat_smooth(method = "lm")

Scales

  • scales are used to modify axes and colors

  • For example:

    • scale_y_continuous() Set name, breaks, labels, limits of y-axis
    • scale_x_log10() log transform the x-axis
    • scale_colour_manual() Specify colors for geoms
    • scale_fill_discrete() Specify colors for geoms

Scales

  • log x and log y axes
ggplot(data = state, aes(x = Population, y = Area, color = state.region)) + scale_y_log10() + scale_x_log10() + geom_point()

Scales

  • Add color manually
ggplot(data = state, aes(x = Population, y = Area, color = state.region)) + scale_y_log10() + scale_x_log10() + geom_point() + scale_colour_manual(values = c("palegreen3","palevioletred3","peachpuff3", "paleturquoise3"))

Labels

  • Add labels, including a title and axes labels, using + labs() 
ggplot(data = state, aes(x = Life.Exp, y = HS.Grad, color = state.region, shape = state.region)) + geom_point(size = 2) + labs(title = "US States, 1977", x = "Life expectancy (in years)", y = "High school graduation rate (percent)")

Labels

  • Add labels inside the figure using + geom_text() 
ggplot(data = state, aes(x = Life.Exp, y = HS.Grad, color = state.region, shape = state.region)) + geom_point(size = 2) + labs(title = "US States, 1977", x = "Life expectancy (in years)", y = "High school graduation rate (percent)") + geom_text(aes(label = state.abb, size = 2, hjust = 0, vjust = 0))

Themes

  • Control over the figure as a whole can be done by modifying themes. See ?theme for all of the options

  • I prefer + theme_bw() (white background with gridlines) or + theme_classic() (white background without gridlines)

  • You can also change the overall font family and font size

ggplot(iris, aes(Species, Sepal.Length)) + geom_bar(stat = "identity") + theme_bw