7 Base R vs. Tidyverse

While my predecessors are Tidyverse zealots, I am not. When it comes to learning R, I don’t feel the need to force anyone into learning it one way or the other. People may disagree with this philosophy! Ultimately, I want you to feel comfortable whenever you encounter R. And, the fact is, many people use both base R and Tidyverse so it is best to know both. In PS 812, for example, you will be using primarily base R functions. So forcing you to learn only Tidyverse at this point feels impractical and incompatible with your broader learning goals.

Throughout this course, I will try to specify when I am using base R and when I am using Tidyverse. There are some functions that I find much easier (and more intuitive) in base R, and others that I prefer written in Tidyverse. But just know this: if there’s a way to write it in Tidyverse, there’s a way to write in base R, and vice versa.

7.1 But first, what’s Tidyverse?

Tidyverse is a R package designed to clean and manipulate data. Compared to base R, Tidyverse can clean and manipulate data much faster with code that is more compact and practical.

However, there are commands that are faster to write and easier to remember in base R, even if it doesn’t process faster than its Tidyverse equivalent. You will figure out which commands those turn out to be.

In general, processing time is dependent on the amount of data you have, so if you only have a few thousand observations, it might not make much of a difference, but it will probably make a difference if you have tens or hundreds of thousands of observations.

7.2 R basics

Some of these exercises were inspired by Mike DeCrescenzo’s R basics materials and Kosuke Imai’s supplementary materials for his book.

Go to RStudio.
Go to File > Open Project in New Session. Find your ps811-exercises folder.
Go to File > New File > R Script.
A new Untitled1 R Script tab will open. Go to File > Save As. Name it lecture5-exercise-basic.R.

7.2.1 Use R as a calculator

Use R to add/subtract/multiply/divide. You don’t need the spaces between each command, but it is easier to read if you do.
```
2 + 3
3 - 1
12 * 12
50 / 2
```
Outputs should be 5, 2, 144, and 25, respectively.
R searches the next line if the inital line is incomplete.
```
50 /
2
```
Output should be 25.
Use R to conduct order of operations.
```
5^(10/5)
```
Output should be 25.
You may also use commands for Euler’s constant and the natural log in R.
```
exp(1)
log(1)
```
R recognizes pi and cosines.
```
pi
cos(pi)
```

If you ever want to figure out how to calculate something in R, just Google it or check out this short Quick-R Operators list for commonly used arithmetic and logical operators.

7.2.2 Creating objects

You want to put 8 + 1 into an object called nine.
```
nine <- 8 + 1
```
See what happens when you type the name of the object, nine. You get an output of 9.
You can also do math with the object. Type in sqrt(nine). You get an output of 3.
Everything can be an object. You can store text in some other text.
```
hello <- "hi"
hello
```
Output should be hi.
If you would like to put something else in the object, you can simply overwrite it.
```
hello <- "hi, nice to meet you, what is your name"
hello
the_answer <- 8 + 1
the_answer
nine <- 50 + 4
nine # the numeric value does not have to be nine
```
Output should be “hi, nice to meet you, what is your name,” 9, and 54, respectively.

7.2.3 Creating vectors

Vectors can be a bunch of different values.

vector_example <- c(5, 6, 7, 8, 9, 10, 100)

You can multiple the vector with an equation, such as 2x.
```
results <- 2 * vector_example
results
```
Output should be 10, 12, 14, 16, 18, 20, 200.
You can multiply vectors.
```
vector_example * results
```
Output should be 50, 72, 98, 128, 162, 200, 20000.
You can do a matrix multiplication using %*%.
```
vector_example %*% results
```
Output should be 20710.
You can plug in the vectors into various functions.
```
mean(vector_example)
median(vector_example)
mode(vector_example)
sd(vector_example) # standard deviation
sum(vector_example)
```
Ouptut should be 20.71429, 8, “numeric” (there is no mode), 25.0034, 145, respectively.

7.3 Loading data into R

When you first start with R, most tutorials will suggest that you use setwd() to call in your directory. And sure, that’s fine. But since this course emphasizes replication, what if someone opens your .Rproj in hopes of replicating your work and get an error when they try to run setwd("/Users/marcyshieh/marcyshieh.github.io/_ps811")? They are not you, they do not work on your computer, so they need to go through the extra step of changing that line.

The person trying to replicate your project can probably type getwd() and replace whatever you have in setwd() with the resulting URL, but you want to make their lives (and your life!) easier.
Consider the here package. The here package allows anyone to replicate your work without being in your exact workspace and directory.
Open a new file. Name it lecture5-exercise-data.R.

Paste the following code into line 1 of your .R file. Highlight all the lines and click “Run” on RStudio OR command+return on MacOS or ctrl+enter on Windows.

# install the here package
install.packages("here")

# load the here library
library("here")
# this is an equivalent to setwd()

# check out where "here" is
here()
# this is an equivalent to getwd()

# as you have probably guessed by now, the "#"" denotes comments and R leaves them alone when you run chunks of code

Download the movie_metadata.csv dataset. This dataset is from Kaggle and contains metadata about movies from IMDb.
Move the movie_metadata.csv from the download location to your ps811-exercises directory.
Go to your .R file. Load the CSV file into your R environment.
```
movie_metadata <- read.csv(here("movie_metadata.csv"))
```
You are calling here() inside the read.csv() command because you are using the here() package.

7.4 Comparing base R vs. Tidyverse

Much of the lesson here is indebted to Hugo Taraves, who wrote a nifty guide to syntax equivalents of base R vs. Tidyverse.

Prerequisites You need to install the dplyr and tidyr packagess, and load them. Use the install.packages() and library() functions to do this. You need these packages to use Tidyverse functions.

Always load your packages on the top of the .R file.

Take a look at the variables in the dataset using names(movie_metadata).

We are going to first go through everything in base R, then repeat everything in Tidyverse. Hopefully this gives you a sense of how base R and Tidyverse work.

7.4.1 Base R

This is how these commands work in base R.

Extract the first 100 rows.
```
movie_metadata_100 <- movie_metadata[1:100, ]
```
- Rows are equivalent to observations.
- Columns are equivalent to variables.

Look at the existing variables in dataset.

# you can identify the variables by name
movie_metadata_100[, c("director_name", "movie_title")] 

# or you can identify the variables by their column index/number
movie_metadata_100[, c(2, 12)]

Create a new variable in the dataset.

# you want to get the sum of facebook likes for the first 3 actors listed in the movie
movie_metadata_100$main_actors_fb_popularity <-
movie_metadata_100$actor_1_facebook_likes +
movie_metadata_100$actor_2_facebook_likes +
movie_metadata_100$actor_3_facebook_likes

# view the dataset then scroll to the last column on the right to see the column you have just created!
View(movie_metadata_100)

Filter rows in dataset.

# you want to extract the observations where there is a IMDB score of 5+ and 2 number of faces in the poster
subset(movie_metadata_100, imdb_score > 5 &
facenumber_in_poster==2)

Arrange rows.

# descending order of movie title followed by ascending order of budget
movie_metadata_100[order(rev(movie_metadata_100$movie_title),
movie_metadata_100$budget) , ]

Summarize observations.

# Create a dataframe with mean and standard deviation information
# na.rm = TRUE removes all the NAs from the calculations
# in this case, if you include the NAs, it will yield a result of NA
# e.g., mean(movie_metadata_100$budget) # not what you want!

data.frame(budget.mean = mean(movie_metadata_100$budget, na.rm = TRUE),
budget.sd = sd(movie_metadata_100$budget, na.rm = TRUE),
gross.mean = mean(movie_metadata_100$gross, na.rm = TRUE),
gross.sd = sd(movie_metadata_100$gross, na.rm = TRUE))

Group observations.

Summarize rows within groups

# Using aggregate
aggregate(formula = cbind(budget, gross) ~ country + genres, 
     data = movie_metadata_100, 
     FUN = function(x){
       c(mean = mean(x), sd = sd(x))
     })

7.4.2 Tidyverse

You know how to do it in base R, so now you can see how it works in Tidyverse! Tidyverse is a suite of packages that are extremely useful for data manipulation. We will have a longer discussion of using tidyverse package for data manipulation next week but hopefully this introduction shows how tidyverse tools deviate from base R tools.

Extract the first 100 rows.
```
movie_metadata_100tidy <- movie_metadata %>% top_n(100)
```
As you have probably noticed, Tidyverse uses %>%, which are called “pipes.” In essence, you are calling your dataset, and piping functions into it.

Look at existing variables in dataset.

# you can identify the variables by name
select(movie_metadata_100tidy, director_name, movie_title)

# or you can identify the variables by their column index/number
select(movie_metadata_100tidy, 2, 12)

If you want to turn this selection into an object, just put movie_metadata_100tidy_select<- (or whatever would make sense for you) in front of the command.

Create a new variable in the dataset.

# you want to get the sum of facebook likes for the first 3 actors listed in the movie

movie_metadata_100tidy <- mutate(movie_metadata_100tidy,
main_actors_fb_popularity = actor_1_facebook_likes +
actor_2_facebook_likes + actor_3_facebook_likes)

# you should have created a brand new variable!
View(movie_metadata_100tidy)

Notice that you don’t have to refer to the data set every time you call a variable. That’s one of the advantages of Tidyverse!

Filter rows in dataset.

# you want to extract the observations where there is a IMDB score of 5+ and 2 number of faces in the poster
filter(movie_metadata_100tidy, imdb_score > 5 & facenumber_in_poster==2)

You may also turn this filtered dataset into an object if you want!

Arrange rows.

# descending order of movie title followed by ascending order of budget
arrange(movie_metadata_100tidy, desc(movie_title), budget)

Summarize observations.

# Create a dataframe with mean and standard deviation information
summarise(movie_metadata_100tidy,
          budget.mean = mean(budget, na.rm = TRUE),
          budget.sd = sd(budget, na.rm = TRUE),
          gross.mean = mean(gross, na.rm = TRUE),
          gross.sd = sd(gross, na.rm = TRUE))

Group observations.

Summarize rows within groups

movie_metadata_100tidy %>% 
     group_by(country, genres) %>% 
     summarise(budget.mean = mean(budget),
       budget.sd = sd(budget),
       gross.mean = mean(gross),
       gross.sd = sd(gross)) %>% 
       ungroup()

 # ungroup() removes any grouping changes you make from analyses beyond this section