In jr-packages/jrTidyverse2: Jumping Rivers: Next Steps in the Tidyverse

Question 1

Load in the files dataset from jrTidyverse2.

library("tidyverse")
data(files, package = "jrTidyverse2")

How many files end in .r? (Hint: Remember that the dot is a special character).

str_subset(files, "\\.r$")

How would you find files which end in .r or .R?

str_subset(files, "\\.(r|R)$")

How many files have atleast one number in them?

str_subset(files, "[0-9]")

Question 2

We'll start off by loading the data

library("tidyverse")
data(beer, package = "jrTidyverse2")

Let's inspect the data

head(beer)

Here we have a data set of beers with their alcohol percentage and colour. Colour is ranked from 1-50 with 1 being pale and 50 being black. The only problem is that the names of the beers have been scraped from the internet and so are contained in an url. To do any analysis on the beers we are going to need to extract the names. The names of the beers are always after the last forward slash in the url. For example, the first url,
/homebrew/recipe/view/61925/the-devil-is-in-the-details-duvel-clone-

would become
The Devil Is In The Details Duvel Clone

It's going to be a bit easier to extract the vector of urls, work with it that way, then reattach it once we are done.

url = beer$URL

1) Extract the last part of the url.
Hint: Your regex should start with / and should end with a $.

url = url %>% 
  str_extract("/[a-zA-Z-_0-9]*$") 

2) Going with the first example, your beer name should now look like
/the_devil_is_in_the_details_duvel_clone_
Get rid of the forward slash.

url = url %>% 
  str_replace("/", "") 

3) The beer names should now look like
the_devil_is_in_the_details_duvel_clone_
Replace the underscores with spaces. Careful, some of the urls have dashes instead of underscores inbetween words.
Hint: Use a group, (), in your regex

url = url %>% 
  str_replace_all("(\\-|\\_)", " ")

# nolint start

4) The beer names will now look like
the devil is in the details duvel clone

Trim the surrounding whitespace and give all words capital letters. Once that is complete, overwrite the urls with the extracted names within the data.

url = url %>% 
  str_to_title() %>% 
  str_trim()
beer$URL = url

# nolint end

5) We want to do some analysis on the beers based on whether they are an IPA, stout or pale ale. To do this we're going to introduce a new function called if_else() from dplyr. For example

df = data.frame(x = c(2, 4, 6, 8))

Here we have made a data frame called df containing a column of numbers called x.

df = df %>% 
  mutate(y = if_else(condition = x > 5, true = 1, false = 0))
df

In this step we are mutating a new column called y that will be the value 1 when x > 5 and the value 0 otherwise. We can do the same for the beers. Notice that if we run the code

str_detect(beer$URL, "Ale")

We get a TRUE when the beer name contains Ale and FALSE otherwise. We can use this inside if_else() as a condition

beer = beer %>% 
  mutate(
    Ale = if_else(str_detect(URL, "Ale"), 1, 0)
         )

So here we would be creating a column called Ale, that is 1 when the beer name contains Ale and 0 otherwise. Create a column called Ipa that is 1 when the name contains Ipa and 0 otherwise. Do the same for Stout.

beer = beer %>% 
  mutate(Ale = if_else(str_detect(URL, "Ale"), 1, 0),
         Ipa = if_else(str_detect(URL, "Ipa"), 1, 0),
         Stout = if_else(str_detect(URL, "Stout"), 1, 0)
         )

6) Under the principles of tidy data, this is no longer tidy, we should have one column containing whether the beer is an "Ale", "Ipa" or "Stout". We can do this using gather() from tidyr

beer = beer %>% 
  gather(Type, Yes, Ale, Ipa, Stout) %>% 
  filter(Yes != 0) %>% 
  select(-Yes)

Here we are gathering the Ale, Ipa and Stout columns into two columns called Type and Yes. We're not interested in the beers with a value of 0 so we filter them out. Then we delete the Yes column using select(). Which type of beer has the highest average alcohol percentage and color? Hint: Use dplyr

7) Plot the data using ggplot2, assigning a different colour to each type of beer

ggplot(beer, aes(x = ABV, y = Color)) + 
  geom_point(aes(colour = Type))

jr-packages/jrTidyverse2 documentation built on Sept. 24, 2020, 4:36 p.m.