Lecture 1 – Introduction to R

Martin Morgan

Contents

1 History

• Statistical programming language
• ‘Free’ software: no cost, open source, broad use
• Extensible: packages (>11,000 on CRAN, >1300 on Bioconductor)
• Key features
  • Intrinsic statistical concepts
  • Vectorized computation
  • ‘Old-school’ scripts rather than graphical user interface – great for reproducibility!
  • (Advanced) copy-on-change semanatics

2 Vectors and data frames

1 + 2

## [1] 3

x = c(1, 2, 3)
1:3             # sequence of integers from 1 to 3

## [1] 1 2 3

x + c(4, 5, 6)  # vectorized

## [1] 5 7 9

x + 4           # recycling

## [1] 5 6 7

Vectors

• numeric(), character(), logical(), integer(), complex(), …
• NA: ‘not available’
• factor(): values from restricted set of ‘levels’.

Operations

• numeric: ==, <, <=, >, >=, …
• logical: | (or), & (and), ! (not)
• subset: [, e.g., x[c(2, 3)]
• assignment: [<-, e.g., x[c(1, 3)] = x[c(1, 3)]
• other: is.na()

Functions

x = rnorm(100)
y = x + rnorm(100)
plot(x, y)

• Many!

data.frame

df <- data.frame(Independent = x, Dependent = y)
head(df)

##   Independent  Dependent
## 1   0.1709466 -1.9343537
## 2  -1.0432251 -1.4366633
## 3   0.1874114 -0.6666756
## 4  -1.0410644 -1.9896486
## 5   0.7765771  1.9752691
## 6  -1.4374832 -1.6424475

df[1:5, 1:2]

##   Independent  Dependent
## 1   0.1709466 -1.9343537
## 2  -1.0432251 -1.4366633
## 3   0.1874114 -0.6666756
## 4  -1.0410644 -1.9896486
## 5   0.7765771  1.9752691

df[1:5, ]

##   Independent  Dependent
## 1   0.1709466 -1.9343537
## 2  -1.0432251 -1.4366633
## 3   0.1874114 -0.6666756
## 4  -1.0410644 -1.9896486
## 5   0.7765771  1.9752691

plot(Dependent ~ Independent, df)  # 'formula' interface

• List of equal-length vectors
• Vectors can be of different type
• Two-dimensional subset and assignment
• Column access: df[, 1], df[, "Indep"], df[[1]], df[["Indep"]], df$Indep

Exercise: plot only values with Dependent > 0, Independent > 0

1. Select rows

   ridx <- (df$Dependent > 0) & (df$Independent > 0)

2. Plot subset

   plot(Dependent ~ Independent, df[ridx, ])

   

3. Skin the cat another way

   plot(
       Dependent ~ Independent, df,
       subset = (Dependent > 0) & (Independent > 0)
   )

   

3 Analysis: functions, classes, methods

fit <- lm(Dependent ~ Independent, df)  # linear model -- regression
anova(fit)                              # summary table

## Analysis of Variance Table
## 
## Response: Dependent
##             Df  Sum Sq Mean Sq F value    Pr(>F)    
## Independent  1 118.609 118.609  118.03 < 2.2e-16 ***
## Residuals   98  98.483   1.005                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

plot(Dependent ~ Independent, df)
abline(fit)

• lm(): plain-old function
• fit: an object of class “lm”
• anova(): a generic with a specific method for class “lm”

class(fit)

## [1] "lm"

methods(class="lm")

##  [1] add1           alias          anova          case.names    
##  [5] coerce         confint        cooks.distance deviance      
##  [9] dfbeta         dfbetas        drop1          dummy.coef    
## [13] effects        extractAIC     family         formula       
## [17] hatvalues      influence      initialize     kappa         
## [21] labels         logLik         model.frame    model.matrix  
## [25] nobs           plot           predict        print         
## [29] proj           qr             residuals      rstandard     
## [33] rstudent       show           simulate       slotsFromS3   
## [37] summary        variable.names vcov          
## see '?methods' for accessing help and source code

4 Help!

?"plot"          # plain-old-function or generic
?"plot.formula"  # method

5 Packages

library(ggplot2)
ggplot(df, aes(x = Independent, y = Dependent)) +
    geom_point() + geom_smooth(method = "lm")

• General purpose: >11,000 packages on CRAN
• Gain contributor’s domain expertise and weird (or other) idiosyncracies
• Installation (once only per computer) versus load (via library(ggplot2), once per session)