Base R vs tidyverse

First of all, this write up is mean for a beginner in R.

Things can be done in many ways in R. In facts, R has been very flexible in this regard compared to other statistical softwares. Basic things such as selecting a column, slicing a row, filtering a data based on certain condition can be done using a base R function. However, all these things can also be done using a tidyverse approach.

Tidyverse basically, a collection of packages that can be loaded in a line of function.

library(tidyverse)

Tidyverse is developed by “RStudio people” pioneered by Hadley Wickham, which means that these packages will be continuously maintained and updated.

So, without further ado, these are the comparisons between these two approaches for some very basic thingy:

1. Select or deselect a column and a row

# Base R
iris[1:5, c("Sepal.Length", "Sepal.Width")]
iris[1:5,c(1,2)] # similar to above
iris[1:5, -1]

# Tidyverse
iris %>% 
  select(Sepal.Length, Sepal.Width) %>% 
  slice(1:5)
iris %>% 
  select(-Sepal.Length) %>% 
  slice(1:5)

2. Filter based on condition

# Base R
iris[iris$Species == "setosa", ]

# Tidyverse
iris %>% 
  filter(Species == "setosa")

3. Mutate (transmute replace the variable)

# Base R
iris$SL_minus10 <- iris$Sepal.Length - 10

# Tidyverse
iris %>% 
  mutate(SL_minus10 = Sepal.Length - 10)

4. Sort variable

# Base R
iris[order(-iris$Sepal.Width),]

# Tidyverse
iris %>% 
  arrange(desc(Sepal.Length))

5. Group by (and get mean for variable Sepal.Width)

# Not really base R
doBy::summaryBy(Sepal.Width~Species, iris, FUN = mean) 

# Tidyverse
iris %>% 
  group_by(Species) %>% 
  summarise(mean_SW = mean(Sepal.Width))

6. Rename variable

# Base R
colnames(iris)[6] <- "hanis"

# Tidyverse
iris %>% 
  rename(Species = hanis)

So, that’s it. Overall, tidyverse give a clarity in understanding the code as it reads from left to right. On the contrary, the base R approach reads from inside to outside, especially for a more complicated code.