Spoon-Fed R

1. interacting with R
2. using R as a calculator
3. variables
4. data structures
5. summarizing data
6. loops, flow-control
7. apply
8. basic stats in R
9. reading and writing delimited data
10. plotting with base R graphics
11. loading and installing packages
12. plotting with ggplot2

• command-line interpreter
• GUI interpreter: RStudio

• everyone has one
• just type R at your command-line shell:

R version 3.0.2 -- "Frisbee Sailing"
Platform: x86_64-apple-darwin13.0.0 (64-bit)
...

Type 'q()' to quit R.

> 

• The carat (>) is your prompt for entering commands
• I will omit the carat for the rest of the presentation

• Download from http://www.rstudio.com/

RStudio is an integrated development environment including:

• interpreter with code completion
• text editor with syntax highlighting and completion
• file browser
• version control manager
• visual object workspace
• command history

# This is a comment, which is ignored

# functions are applied with ()
print("hello") 

[1] "hello"

• anything in quotes is a “string”
• anything else is either a number or:
  • function
  • class
  • operator (+-/?%&=<>|!^*)

Addition

2 + 2

[1] 4

Subtraction

5 - 2

[1] 3

Division

2 * 2

[1] 4

Multiplication

5 / 2

[1] 2.5

Exponents

2^4

[1] 16

Logorithms

log10(100)

[1] 2

log2(4)

[1] 2

Order of operations

10 / 2 - 1

[1] 4

10 - 5 / 5

[1] 9

(10 - 5) / 5

[1] 1

Be careful. Evaluation of operators occurs left to right.

x <- 1
x

[1] 1

Variables can be assigned (<-) a value

x <- 1
y <- 2
x <- y
x

[1] 2

y

[1] 2

But be careful because they can be re-assigned

x <- 0

x > 1 ## x is greater than 1

[1] FALSE

x < 1 ## x is greater than 1

[1] TRUE

x == 1

[1] FALSE

x == 0

[1] TRUE

x != 0

[1] FALSE

Comparisons result in boolean values

x <- 3
y <- c(1,2,x)
y

[1] 1 2 3

Vectors can hold elements of the same type.

names(y) <- c("one", "two", "three")
y

  one   two three 
    1     2     3 

Vectors can also have names for each element.

z <- y * 3
z

  one   two three 
    3     6     9 

sum(z)

[1] 18

Arithmetic can be performed on a vector, which applies that operation to every element and returns a new vector.

  one   two three 
    3     6     9 

z[1]

one 
  3 

z["one"]

one 
  3 

Vectors can be indexed using a 1-based position, as well as name.

z

  one   two three 
    3     6     9 

z[2:3]

  two three 
    6     9 

Slicing a vector is as easy as specifying start:end.

z[-1]

  two three 
    6     9 

z[-2:-3]

one 
  3 

Remove elements from a vector using negative indices.

q <- list(y, z)
q

[[1]]
  one   two three 
    1     2     3 

[[2]]
  one   two three 
    3     6     9 

Lists can contain vectors.

q[[1]]

  one   two three 
    1     2     3 

q[[1]][1]

one 
  1 

You can index a list in the same way as a vector.

v <- seq(1,9) ## or 1:9
v

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

Let's construct a sequence of 9 numbers.

c(v,v)

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

rep(v, times=3)

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

We can concatonate or repeat a vector as well.

mt <- matrix(v, nrow=3)
mt

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

matrix(v, nrow=3, byrow=T)

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

Matrices, created from vectors, are row or column oriented.

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

mt[1,1]

[1] 1

mt[3,3]

[1] 9

Matrices are indexed as [row,col]

dim(mt)

[1] 3 3

nrow(mt)

[1] 3

ncol(mt)

[1] 3

Dimensionality, number of rows and columns can computed using these functions.

df <- data.frame(y, z)
colnames(df) <- c("first","second")
df

      first second
one       1      3
two       2      6
three     3      9

Dataframes are like matrices, but contain more structure.

      first second
one       1      3
two       2      6
three     3      9

df$first

[1] 1 2 3

Dataframes can be indexed by name to return a vector.

      first second
one       1      3
two       2      6
three     3      9

df["first"]

      first
one       1
two       2
three     3

Dataframes can be indexed by name to return another dataframe

      first second
one       1      3
two       2      6
three     3      9

df$first[1]

[1] 1

Dataframes can be further indexed to return individual elements

      first second
one       1      3
two       2      6
three     3      9

df > 3

      first second
one   FALSE  FALSE
two   FALSE   TRUE
three FALSE   TRUE

Dataframes, just like other structures, can be compared, resulting a boolean values.

      first second
one   FALSE  FALSE
two   FALSE   TRUE
three FALSE   TRUE

df[df > 3]

[1] 6 9

Passing the boolean result of comparison as an index returns only elements where the comparison was TRUE.

      first second
one   FALSE  FALSE
two   FALSE   TRUE
three FALSE   TRUE

which(df > 3)

[1] 5 6

The which function converts a boolean index to a numeric index.

      first second
one       1      3
two       2      6
three     3      9

cbind(df, data.frame("third"=c(9,18,27)))

      first second third
one       1      3     9
two       2      6    18
three     3      9    27

Dataframe columns can be bound to form a new dataframe.

      first second
one       1      3
two       2      6
three     3      9

rbind(df, data.frame("first"=4, "second"=12, row.names="four"))

      first second
one       1      3
two       2      6
three     3      9
four      4     12

Dataframe rows can be bound to form a new dataframe.

library(datasets)
dim(cars)

[1] 50  2

head(cars)

  speed dist
1     4    2
2     4   10
3     7    4
4     7   22
5     8   16
6     9   10

mean(cars$speed)

[1] 15.4

median(cars$speed)

[1] 15

sd(cars$speed)

[1] 5.288

Mean, median and standard deviation.

summary(cars)

     speed           dist    
 Min.   : 4.0   Min.   :  2  
 1st Qu.:12.0   1st Qu.: 26  
 Median :15.0   Median : 36  
 Mean   :15.4   Mean   : 43  
 3rd Qu.:19.0   3rd Qu.: 56  
 Max.   :25.0   Max.   :120  

Summarizing a dataframe returns percentiles and mean.

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

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

Use for loops to repeat a task a certain number of times.

x <- 0
if (x == 0) { print("yes") }

[1] "yes"

if (x > 1) { print("yes") } else { print("no") }

[1] "no"

• If statements only execute code if the condition evaluates to TRUE.
• Else statements execute when the condition is not satisfied.

x <- 0
while (x < 5){ 
  print(x) 
  x <- x + 1
  }

[1] 0
[1] 1
[1] 2
[1] 3
[1] 4

Use while loops to repeat a task while a condition (x<5) is true.

      first second
one       1      3
two       2      6
three     3      9

apply(df, 1, sum)

  one   two three 
    4     8    12 

apply(df, 2, sum)

 first second 
     6     18 

Apply a function over array columns (1) or rows (2).

      first second
one       1      3
two       2      6
three     3      9

sapply(df, sqrt)

     first second
[1,] 1.000  1.732
[2,] 1.414  2.449
[3,] 1.732  3.000

Simple apply a function to every element, returning the same type of data structure.

write.table(df, file = "example.txt")
write.table(df, file = "example.tsv", sep = "\t")
write.csv(df, file = "example.csv")

Write 1) space-delimited, 2) tab-delimited, 3) comma-delimited files containing dataframe df.

df1 = read.table("example.txt", header=T)
df2 = read.delim("example.tsv", sep = "\t")
df3 = read.csv("example.csv", row.names = 1)

identical(df1,df2)

[1] TRUE

identical(df2,df3)

[1] TRUE

All three files result in equivalent dataframes.

Issues to consider when reading and writing delimited files:

1. Do I want/have column names (header)?
2. Do I want/have row names?
3. What is my delimiter?
4. Do I want/have quotes surrounding each value?

Check the default behavior of the reading/writing function first.

head(cars)

  speed dist
1     4    2
2     4   10
3     7    4
4     7   22
5     8   16
6     9   10

lmcars <- lm(dist ~ speed, cars)
lmcars

Call:
lm(formula = dist ~ speed, data = cars)

Coefficients:
(Intercept)        speed  
     -17.58         3.93  

• lm fits a linear model: response ~ terms
• in this case the response is distance traveled at speed