R packages

For this activity you will need the dplyr and tidyr packages. Each of these packages is part of the tidyverse, a collection of R packages with a common philosophy for data analysis. If you don’t have the tidyverse package installed, go to your R console and install it first with install.packages("tidyverse") (remember you only need to do this once).

Once the tidyverse package is installed, load it into R with library(tidyverse).

Data

In this activity we will work with data on 344 penguins recorded near Palmer Station, Antarctica. Variables include

• species: penguin’s species (Adelie, Chinstrap, Gentoo)
• island: island where penguin measured (Biscoe, Dream, Torgersen)
• bill_length_mm: penguin’s bill length (mm)
• bill_depth_mm: penguin’s bill depth (mm)
• flipper_length_mm: penguin’s flipper length (mm)
• body_mass_g: penguin’s body mass (g)
• sex: penguin’s sex (female, male)
• year: year when data recorded (2007, 2008, 2009)

This information is contained in the penguins dataset, which is part of the palmerpenguins R package. If you don’t have the palmerpenguins package installed, go to your R console and install it first with install.packages("palmerpenguins") (remember you only need to do this once).

Once the palmerpenguins package is installed, load it into R with library(palmerpenguins).

Looking at the data

To begin, let’s look at our data. The glimpse function is useful for taking a peek at a dataset.

glimpse(penguins)

What does this output tell us?

• The number of rows (i.e., the number of penguins): 344
• The number of columns (i.e., the number of variables recorded for each penguin): 8
• The names of each variable (e.g., species, island, etc.), and the first few observations in each variable (Adelie, Torgersen, etc.)

Now look at the bill length for the fourth penguin. Instead of a number, you get NA. What does NA mean?

In R, NA stands for “Not Available”, and it means that this value is missing in our data. Missing data can be a problem, because R doesn’t know how to handle missing values when we calculate summary statistics or fit models.

Handling missing data

A simple way of dealing with missing data is to remove any rows which contain missing values. How do we do this?

penguins <- penguins %>%
  drop_na()

What’s going on in this code? First, we take the penguins data:

penguins

Next, we want to remove missing values:

penguins %>%
  drop_na()

• The drop_na() function says “remove any rows with missing values”
• The %>% means “Take <THIS>, then do <THAT>”. So penguins %>% drop_na() means “take penguins, then remove rows with missing values”

Finally, we need to save the result. The <- is like “Save As…”, and means “take what is on the right hand side, and save it as the left hand side”. So penguins <- penguins %>% drop_na() means “modify penguins to remove rows with missing values”.

glimpse(penguins)

Making a subset of data

We have three species of penguin: Adelie, Chinstrap, and Gentoo. Ultimately, we might want to compare characteristics between these different species, but having groups in the data can make initial exploration more challenging. A good strategy is to begin by focusing on just one group. In this case, we will focus on just one species of penguin for now.

Let’s focus on the Chinstrap penguins. How do we do that? We can use the filter function, which keeps only the rows which satisfy a specified condition.

chinstrap_penguins <- penguins %>%
  filter(species == "Chinstrap")

What’s going on in this code? Remember that the pipe %>% means “take <THIS>, then do <THAT>”. So we take the penguins data, then filter so that only the Chinstrap penguins are left (species == "Chinstrap"). Finally, we save the result as a new dataset, which we call chinstrap_penguins.

• Note that we use two equals signs (==) when we are checking whether species is Chinstrap
• Because Chinstrap is a word (rather than a number), we need to put it in quotes ("Chinstrap") for R to interpret the code correctly.

glimpse(chinstrap_penguins)

Summary statistics

Now that we have a subset of data, let’s explore some variables! One part of exploring variables is summary statistics, like the mean, standard deviation, median, and IQR. Let’s start with the mean bill length, which is calculated with the mean function in R.

chinstrap_penguins %>%
  summarize(mean_bill_length = mean(bill_length_mm))

This tells us that the mean bill length for Chinstrap penguins in the data is 48.8 mm. How does the code work?

• The summarize function mean “calculate summary statistics”
• Inside the summarize function, we calculate the statistics we want
• In this case, we want the mean bill length (mean(bill_length_mm))
• Finally, we give our summary statistic a meaningful name: mean_bill_length (we could call it whatever we want)

chinstrap_penguins %>%
  summarize(mean_body_mass = mean(body_mass_g))

chinstrap_penguins %>%
  summarize(median_body_mass = median(body_mass_g))

adelie_penguins <- penguins %>%
  filter(species == "Adelie")
adelie_penguins %>%
  summarize(mean_bill_length = mean(bill_length_mm))

Comparing groups

In Question 6, you calculated the mean bill length for Adelie penguins. But to do that, we had to first subset the data to pull out the Adelies. Is there a way to compare statistics between groups, without lots of subsetting?

Fortunately, there is! The group_by function allows us to group our data before calculating summary statistics.

penguins %>%
  group_by(species) %>%
  summarize(mean_bill_length = mean(bill_length_mm))

Now we get the average bill length for each species. What’s going on in this code?

• First, notice that we can chain our pipes %>% together. The output of one line is the input to the next line. This code means “take our penguins, THEN group by species, THEN calculate summary statistics”
• The group_by function creates groups based on the values of one or more variables. Here, our groups are defined by the species variable, so we get one group per species. After grouping, summary statistics are calculated separately for each group.

penguins %>%
  group_by(species) %>%
  summarize(mean_body_mass = mean(body_mass_g))

penguins %>%
  group_by(sex) %>%
  summarize(mean_body_mass = mean(body_mass_g))

penguins %>%
  group_by(species, sex) %>%
  summarize(mean_body_mass = mean(body_mass_g))

Creating new variables

Finally, in exploring our data we may want to create new variables. For example, suppose we care about the ratio of body mass to flipper length. We can use the mutate function to create new variables.

penguins <- penguins %>%
  mutate(bf_ratio = body_mass_g/flipper_length_mm)

What’s going on in this code?

• mutate creates a new variable
• We call this new variable bf_ratio
• bf_ratio is defined as body_mass_g/flipper_length_mm

penguins %>%
  group_by(species) %>%
  summarize(median_bf = median(bf_ratio))

penguins <- penguins %>%
  mutate(bill_diff = bill_length_mm - bill_depth_mm)
penguins %>%
  group_by(species) %>%
  summarize(max_diff = max(bill_diff))