Examples from lesson #1 - Intro to R - presented by Diego Barneche

Vector types

There are 5 different types of vector in R.

x <- TRUE       # x becomes TRUE
x <- 1          # x now has a value of 1 (overwritten the line above)
x <- c(1, TRUE) # x now contains two values of 1, because a vector always stores the same mode of elements
x

## [1] 1 1

x <- c(1, TRUE, '1') #now all values are characters, because there is a underlying hierarchy. For instance, if there is one element of mode 'character' in your vector, then all elements will be treated as so
x

## [1] "1"    "TRUE" "1"

Lists

If you want to store the different modes of elements in one single object, you use lists.

x <- list(1, TRUE, "1")
x

## [[1]]
## [1] 1
## 
## [[2]]
## [1] TRUE
## 
## [[3]]
## [1] "1"

Querrying

You can ask yourself about the nature of an unknown vector - suppose you did not know what x is:

storage.mode(x)

## [1] "list"

typeof(x)

## [1] "list"

class(x)

## [1] "list"

Examples from lesson #2 - Writing functions - presented by Diego Barneche

Playing with the basics

When you write your function, you define it based on arguments

f <- function(something) {
    something
}
f(100)

## [1] 100

The function f() is what we call the identity function, i.e., the function returns the exact input. In this particular case, the argument something took the value 100. You could instead define an object before, and then calling the function on that object, which should do exactly the same thing.

x  <-  100 #x now is a numeric vector of one element of value 100
f(x)

## [1] 100

It returned exactly the same thing. In this case, something took the value of x, which was 100.

Downloading the gapminder data

We have to download a .csv file that we will use for the rest of our course. Right-click this link and save it to some new directory. Let’s play with it now.

data  <-  read.csv('../data/gapminder-FiveYearData.csv', header=TRUE, stringsAsFactors=FALSE)
head(data) #try the function tail() instead, what happened?

##       country year      pop continent lifeExp gdpPercap
## 1 Afghanistan 1952  8425333      Asia  28.801  779.4453
## 2 Afghanistan 1957  9240934      Asia  30.332  820.8530
## 3 Afghanistan 1962 10267083      Asia  31.997  853.1007
## 4 Afghanistan 1967 11537966      Asia  34.020  836.1971
## 5 Afghanistan 1972 13079460      Asia  36.088  739.9811
## 6 Afghanistan 1977 14880372      Asia  38.438  786.1134

#subset the data for year 1982 only
data.1982  <-  data[data$year == 1982, ]

#let's create a function to compute the arithmetic average of a vector
average  <-  function(x) {
    sum(x) / length(x)
}

average(data.1982$gdpPercap)

## [1] 7518.902

#same as 
sum(data.1982$gdpPercap) / length(data.1982$gdpPercap)

## [1] 7518.902

Notice that repeating things may be problematic; first because it increases your chances of typos; second, if for some reason you change the name of your columns in dat, you will have to manually change all these names, which will be quite annoying.

Examples from lesson #5 - Testing functions - presented by Diego Barneche

We have seen how important it is to test your functions in R. We have used a very user-friendly packages called testthat, where most tests are expectations of a certain result/output:

Example of test

#using function rescale as an example
rescale <- function(x, r.out) {
  if (length(r.out) != 2)
    stop("Expected r.out to be length 2")
  xr <- range(x, na.rm=TRUE)
  p <- (x - min(xr)) / (max(xr) - min(xr))
  r.out[[1]] + p * (r.out[[2]] - r.out[[1]])
}

#some of the possible tests for it
library(testthat)
test_that("Rescale gives expected range", {
  x <- rnorm(20)
  x[4] <- NA
  r.out <- sort(runif(2))
  ## Range is expected:
  expect_that(range(rescale(x, r.out), na.rm=TRUE), equals(r.out))
  ## Rescaling onto same range does not change the data:
  expect_that(rescale(x, range(x, na.rm=TRUE)), equals(x))
  
  ## Rescaling onto a reversed range works
  expect_that(range(rescale(x, rev(r.out)), na.rm=TRUE), equals(r.out))
  ## And the output is what was expected:
  expect_that(rescale(x, rev(r.out)),
              equals(sum(r.out) - rescale(x, r.out)))
})

Examples from lessons #2 and #4 - Writing plot functions and looping - presented by Diego Barneche and Raniere Silva

Making a plot

col.table  <-  c(Asia='tomato',
                 Europe='chocolate4',
                 Africa='dodgerblue2',
                 Americas='darkgoldenrod1',
                 Oceania='green4')

for (year in unique(data$year))
    my.plot(year, data, col.table)