• https://www.r-project.org
  • GNU project since 1997
• Name from the authors (Ross Ihaka and Robert Gentleman)
• Influenced by S lang
  • Originally written in Fortran
  • The idea was that you can initially analyse the data without thinking with the “programming” mindset

Rmd

• Tables: https://rmarkdown.rstudio.com/lesson-7.html
• My templates

Resources to learn

• https://swirlstats.com/
• Starting data analysis/wrangling with R: Things I wish I’d been told
• https://r4ds.had.co.nz/data-visualisation.html
• https://r4ds.had.co.nz/graphics-for-communication.html
• https://www.r-graph-gallery.com/

Basic commands

• ls()
  • List all objects
• rm(<v>)
  • remove object
  • rm(list = ls())
    • Removes all elements on the list (in this case, everything)
• class(<v>)
• is.na(<v>)
• is.nan(v>)

Atomic Objects

• Numeric (double)
• Integer
• Complex (complex numbers, e,g. 1 + 4i)
• Logical (boolean)
• Character (actually a String as we know in other languages)

Vectors

• You can use it as a dictionary as well (names)

> x <- c(1,2)
> names(x) <- c("aa", "bb")
> x
aa bb 
 1  2 

> x <- c(1,2,3,4) # concat vectors

Factors

• “Categories”
• Is possible to define the “size” of each level

> factor(c("A", "B", "C"), labels = c("A", "B", "C"), ordered = T)
[1] A B C
Levels: A < B < C

• Usage:

factor(x = character(), levels, labels = levels,
       exclude = NA, ordered = is.ordered(x), nmax = NA)

> factor(c("A", "B", "C"))
[1] A B C
Levels: A B C

Matrixes

• By default they are built by column (needs to set byrow to TRUE if otherwise)

> matrix(data = 1:6, nrow = 2, ncol = 3)
     [,1] [,2] [,3]
[1,]    1    3    5
[2,]    2    4    6

> attributes(matrix( data = 1:6, nrow = 2, ncol = 3))
$dim
[1] 2 3

> x <- c(1,2,3,4)
> dim(x) <- c(2,2)
> x
     [,1] [,2]
[1,]    1    3
[2,]    2    4

• rbind()
• cbind()
• To create names for matrixes, you can use dimnames

List

Data frames

• A special type of list

  • All elements are vectors, each one represents a column. They must have the same number of subelements.

• Different columns can have different types

• You can set stringAsFactores. Prior to R 4.0 it was set as TRUE by default

• data.frame()

• read.table()

• read.csv()

• cbind/rbind can be used with Data Frames as well

Subcollections

• [] return object from the same class as the original object
  • Negative indexes excludes it:

> x <- 1:4; x[-2]
[1] 1 3 4

• You can also use a boolean vector:

 > x <- 1:4; x[x > 2]
[1] 3 4

• You can also use it kinda like a dictionary

> x <- c(1,2); names(x) <- c("aa", "bb"); x["aa"]
aa 
 1 

• For matrixes you can use:

> m[2, ] # second line
> m[, 3] # third column
> m[, 3, drop=FALSE] # third column and returns it as a column

• You can use partial names with $:

> x <- list(foo = 1:4, bar = 1); x$f
[1] 1 2 3 4

Dates

• Date type
• Number of days since 01/01/1970

> unclass(as.Date('1976/01/01'))
[1] 2191

Timestamps

• POSIXct and POSIXlt types

  • POSIXct is just a number, useful to save space
  • POSIXlt is a list with other informations (day of the week, day of the year, month and so on)

• [[]] extract the element of lists and data frames (class of the object and not of its “parent”)

• $ extracts elements based on the name of a list our data frame (similar to [[]])

Curious stuff

• %*% is matrix multiplication
• round uses the ISO rule to round x.5 numbers. If pair it will round to get closer to zero, if it’s odd it will round to get away from zero. (https://en.wikipedia.org/wiki/ISO_80000-1)

> 1/0
[1] Inf

> 0/0
[1] NaN

Random

> 10:20
[1] 10 11 12 13 14 15 16 17 18 19 20

> seq(from=7, by=4, to=20)
[1]  7 11 15 19

• order(x) returns the order of the elements:

> x <- runif(5);
> x
[1] 0.03387367 0.53094936 0.26855677 0.96293228 0.01368555
> order(x)
[1] 5 1 3 2 4

• cor() correlation

Functions

• https://bookdown.org/rdpeng/rprogdatascience/scoping-rules-of-r.html
• Functions always return the last value, like Ruby; (you can also use return())
• function() creates function “objects”
  • f <- function(<args>) { }
  • formals() returns the list of formal arguments
  • Not all arguments are required, even if they dont have a default value
    • You can verify if an argument was given with missing()
  • You can pass arguments by positioning it in the right order or explicitly saying the name
  • You can also use ... arg, if you want to use it you need to either create a vector or a list (e.g. c(...), list(...))
  • All arguments are passed by value (copy)
• To vectorise a function that originally does not know how to handle vectors you can use Vectorize(function)

Good to know

• lapply: apply function to each element of a list, result is a list
• sapply: similar to lapply, but the result is simplified (if the result length is one, the result is a vector, if it’s a list with each element of the same size, the result is a matrix. Otherwise is a list)
• apply: Input is a matrix or a data frame, apply a function to all values of a certain dimension (lines or columns)
• mapply: mapply(log, 2:5, 2:3) (calls log passing an element of first argument and an element of the second argument). The result is also simplified like sapply. From the docs: mapply is a multivariate version of sapply. mapply applies FUN to the first elements of each … argument, the second elements, the third elements, and so on. Arguments are recycled if necessary.
• tapply: Apply a function to each cell of a ragged array, that is to each (non-empty) group of values given by a unique combination of the levels of certain factors. (tapply(mtcars$mpg, mtcars$cyl, mean) = groups the cars by $cyl$ and gets the mean of the $mpg`)

Reading data

• read.csv is just a thin wrapper around read.table.

ggplot2

• https://ggplot2.tidyverse.org/
• You can add “layers” on top of previous layers to make it more rich. From the documentation:

  It’s hard to succinctly describe how ggplot2 works because it embodies a deep philosophy of visualisation. However, in most cases you start with ggplot(), supply a dataset and aesthetic mapping (with aes()). You then add on layers (like geom_point() or geom_histogram()), scales (like scale_colour_brewer()), faceting specifications (like facet_wrap()) and coordinate systems (like coord_flip()).")

• aes() let you define aestetic things on your chart. If you define aes when you call ggplot, the next layears can extend it.