Tidyverse

Introduction

The so called tidyverse is a collection of connected R-packages that are created by the same people that also made R-studio. You have already seen one of these packages: ggplot. Like ggplot, many of these packages do things that could, in principle, also be done in base R but in a may that some people find more intuitive. Because these packages are so and general popular it may well well happen that when you try to find something R-related on the internet you find a solution using the tidyverse. For this reason we give a short introduction to the tidyverse here.

tibble

The package tibble provides the tibble class and some functions to work with this class. A tibble is an alternative to a data.frame and is mostly (but not completely) compatible with this class. The most important differences are the following: Unlike a data.frame a tibble does not use row names. When a single column is selected from a data.frame using [,] the result is converted to a vector. This does not happen with a tibble. Finally, when a tibble is printed only a few rows and columns are shown. tibble is an important package because most other tidyverse packages use tibbles.

dplyr

The dplyr package provides new methods to do data manipulation. It also provides the so-called pipe operator %>% that takes the output of a function on the left-hand-side and uses it as input for a function on the right (by default as the fist argument). This often can make long sequences of transformations much more readable.

library(dplyr)

Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

starwars %>% 
  select(mass, height, gender, homeworld) %>% 
  group_by(gender, homeworld) %>% 
  filter(mass > mean(mass, na.rm = TRUE)) %>% 
  summarise(min_height=min(height,na.rm = TRUE), n=n())

`summarise()` has grouped output by 'gender'. You can override using the
`.groups` argument.

# A tibble: 7 × 4
# Groups:   gender [2]
  gender    homeworld min_height     n
  <chr>     <chr>          <int> <int>
1 feminine  Mirial           170     1
2 masculine Corellia         180     1
3 masculine Kamino           229     1
4 masculine Kashyyyk         234     1
5 masculine Naboo            170     3
6 masculine Tatooine         178     2
7 masculine <NA>             193     2

readr

A package to read text files, mostly much faster than base R does.

purrr

The package purrr makes functional programming with R easier. That is it contains methods that apply functions to the elements of lists or otherwise work with lists and functions. An example is shown below:

library(dplyr)
library(purrr)
data("ToothGrowth")
ToothGrowth %>% 
  split(.$supp) %>% 
  map(function(x) lm(len ~ dose, data=x)) %>% 
  map(coef) %>% 
  bind_rows()

# A tibble: 2 × 2
  `(Intercept)`  dose
          <dbl> <dbl>
1         11.5   7.81
2          3.29 11.7 

forcats

This package makes working with factors easier.

stringr

The stringr package helps you ro work with character data (strings). It does things like merging and splitting and otherwise manipulating character variables and finding and selecting substrings in larger strings.

tidyr

The tidyr package can be used to make data.frames or tibbles ‘tidy’. That is make sure it is organised in a way so that each variable is in a column of its own and each observation has a separate row. This is the format that is expected by most functions in R. However many data sets (for example from excel) do not have this structure. The package also contains functions to reshape data (from long to wide format or vice-versa) or work with nested data, that is tibbles that are stored in single cells of an other table.