01_Rintro.md

title
Introduction to R

Logistics

Presentation{target="_blank"}

R Script{target="_blank"} Download this file and open it (or copy-paste into a new script) with RStudio so you can follow along.

First Steps

Variables

x=1
x

## [1] 1

We can also assign a vector to a variable:

x=c(5,8,14,91,3,36,14,30)
x

## [1]  5  8 14 91  3 36 14 30

And do simple arithmetic:

x+2

## [1]  7 10 16 93  5 38 16 32

Create a new variable called `y` and set it to `15`

Show Solution

y=15

Note that R is case sensitive, if you ask for X instead of x, you will get an error

X
Error: object 'X' not found

Variable naming conventions

Naming your variables is your business, but there are 5 conventions to be aware of:

• alllowercase: e.g. adjustcolor
• period.separated: e.g. plot.new
• underscore_separated: e.g. numeric_version
• lowerCamelCase: e.g. addTaskCallback
• UpperCamelCase: e.g. SignatureMethod

Subsetting

x

## [1]  5  8 14 91  3 36 14 30

Subset the vector using x[ ] notation

x[5]

## [1] 3

You can use a : to quickly generate a sequence:

1:5

## [1] 1 2 3 4 5

and use that to subset as well:

x[1:5]

## [1]  5  8 14 91  3

Using Functions

To calculate the mean, you could do it manually like this

(5+8+14+91+3+36+14+30)/8

## [1] 25.125

Or use a function:

mean(x)

## [1] 25.125

Type ?functionname to get the documentation (?mean) or ??"search parameters (??"standard deviation") to search the documentation. In RStudio, you can also search in the help panel. mean has other arguments too:

mean(x, trim = 0, na.rm = FALSE, ...)

In RStudio, if you press TAB after a function name (such as mean( ), it will show function arguments.

Calculate the standard deviation of `c(3,6,12,89)`.

Show Solution

y=c(3,6,12,89)
sqrt((sum((y-mean(y))^2))/(length(y)-1))

## [1] 41.17038

#or
sd(y)

## [1] 41.17038

#or
sd(c(3,6,12,89))

## [1] 41.17038

Writing functions in R is pretty easy. Let's create one to calculate the mean of a vector by getting the sum and length. First think about how to break it down into parts:

x1= sum(x)
x2=length(x)
x1/x2

## [1] 25.125

Then put it all back together and create a new function called mymean:

mymean=function(f){
  sum(f)/length(f)
}

mymean(f=x)

## [1] 25.125

Confirm it works:

mean(x)

## [1] 25.125

Any potential problems with the `mymean` function?

Missing data: dealing with NA values

x3=c(5,8,NA,91,3,NA,14,30,100)

" What do you think `mymean(x3)` will return?

Calculate the mean using the new function

mymean(x3)

## [1] NA

Use the built-in function (with and without na.rm=T)

mean(x3)

## [1] NA

mean(x3,na.rm=T)

## [1] 35.85714

Writing simple functions is easy, writing robust, reliable functions can be hard...

Logical values

R also has standard conditional tests to generate TRUE or FALSE values (which also behave as 0s and 1s. These are often useful for filtering data (e.g. identify all values greater than 5). The logical operators are <, <=, >, >=, == for exact equality and != for inequality.

  x

## [1]  5  8 14 91  3 36 14 30

  x3 > 75

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

  x3 == 40

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

  x3 >   15

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

And you can perform operations on those results:

sum(x3>15,na.rm=T)

## [1] 3

or save the results as variables:

result =  x3 >  3
result

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

Define a function that counts how many values in a vector are less than or equal (`<=`) to 12.

Show Solution

mycount=function(x){
  sum(x<=12)
}

Try it:

x3

## [1]   5   8  NA  91   3  NA  14  30 100

mycount(x3)

## [1] NA

oops!

mycount=function(x){
  sum(x<=12,na.rm=T)
}

Try it:

x3

## [1]   5   8  NA  91   3  NA  14  30 100

mycount(x3)

## [1] 3

Nice!

Generating Data

There are many ways to generate data in R such as sequences:

seq(from=0, to=1, by=0.25)

## [1] 0.00 0.25 0.50 0.75 1.00

and random numbers that follow a statistical distribution (such as the normal):

a=rnorm(100,mean=0,sd=10)

Let's visualize those values in a histogram:

hist(a)

We'll cover much more sophisticated graphics later...

Data Types

Matrices

You can also use matrices (2-dimensional arrays of numbers):

y=matrix(1:9,ncol=3)
y

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

Matrices behave much like vectors:

y+2

##      [,1] [,2] [,3]
## [1,]    3    6    9
## [2,]    4    7   10
## [3,]    5    8   11

and have 2-dimensional indexing:

y[2,3]

## [1] 8

Create a 3x3 matrix full of random numbers. Hint: `rnorm(5)` will generate 5 random numbers

Show Solution

matrix(rnorm(9),nrow=3)

##             [,1]        [,2]      [,3]
## [1,]  0.05710626 -1.15878801 0.2178735
## [2,]  0.90820862  0.01480023 0.2220378
## [3,] -0.40504985  1.23441876 0.8151335

Data Frames

Data frames are similar to matrices, but more flexible. Matrices must be all the same type (e.g. all numbers), while a data frame can include multiple data types (e.g. text, factors, numbers). Dataframes are commonly used when doing statistical modeling in R.

data = data.frame( x = c(11,12,14),
                   y = c("a","b","b"),
                   z = c(T,F,T))
data

##    x y     z
## 1 11 a  TRUE
## 2 12 b FALSE
## 3 14 b  TRUE

You can subset in several ways

mean(data$x)

## [1] 12.33333

mean(data[["x"]])

## [1] 12.33333

mean(data[,1])

## [1] 12.33333

Loading Packages

For installed packages: library(packagename).

New packages: install.packages() or use the package manager.

library(ggplot2)

R may ask you to choose a CRAN mirror. CRAN is the distributed network of servers that provides access to R's software. It doesn't really matter which you chose, but closer ones are likely to be faster. From RStudio, you can select the mirror under Tools→Options or just wait until it asks you.

If you don't have the packages above, install them in the package manager or by running install.packages("raster").

Today's task

Now complete the first task here by yourself or in small groups.