lab 1 · tidyverse (R) — source
A. Stamm 
This page shows the source file of the lab below. Use the button to download it as a plain file and run it in your own workspace.
# ============================================================
# Lab 1: Tidyverse data transformation: the pipe, dplyr, tidyr
# The R Statistical Language · Session 2
#
# The foundation for everything that follows this session and
# in the Reproducible Workflows course. Comment your thought
# process as you go. Don't peek at a neighbour's tibble; the
# point is to reason about each verb.
# ============================================================
# Load the data once, at the top.
penguins <- readr::read_csv("penguins.csv", show_col_types = FALSE)
## Part A: the pipe and the core verbs ----------------------
# A1. Write a pipe that keeps only the Gentoo penguins.
# penguins |> filter(species == "Gentoo")
# A2. Select three columns that interest you; call the result `gentoo`.
# gentoo <- penguins |> filter(species == "Gentoo") |>
# select(...)
# A3. Add a column `mass_kg = body_mass_g / 1000` to `gentoo`.
# mutate() keeps every other column.
# A4. Sort the Gentoo penguins from lightest to heaviest, then
# heaviest first (arrange() and arrange(desc(...))).
## Part B: summaries by group -------------------------------
# B1. Mean and standard deviation of bill_length_mm per island:
# penguins |> group_by(island) |>
# summarise(avg = mean(bill_length_mm, na.rm = TRUE),
# sd = sd(bill_length_mm, na.rm = TRUE))
# Compare the gentoo and Adelie islands. Which is most variable?
# B2. Which species has the heaviest *median* body mass?
# Same shape as B1, but finish with arrange(desc(median_mass))
# and read the first row. Hint: summarise(median_mass = median(body_mass_g, na.rm = TRUE))
# (na.rm = TRUE: there are missing values.)
# B3. How many individuals of each species live on each island?
# group_by(species, island) |> summarise(n = n())
# Does any species/island combo have no data?
## Part C: reshape with tidyr (the trickiest part; take your time) -----
# C1. Build a WIDE table: average bill length, one row per island and one
# column per species. You already did "summarise per group" in Part B;
# now you just add the pivot step on the end:
# wide <- penguins |> group_by(island, species) |>
# summarise(avg = mean(bill_length_mm, na.rm = TRUE)) |>
# pivot_wider(names_from = species, values_from = avg)
# If C1 feels hard, that's normal; pivot_wider is the newest idea here.
# C2. Melt it back to tidy with pivot_longer():
# wide |> pivot_longer(cols = c(Adelie, Chinstrap, Gentoo),
# names_to = "species", values_to = "avg_bill")
## Part D: reflect -----------------------------------------
# D1. In a comment: for the summary you built in B2, would this be
# easier or harder in base R (tapply / aggregate / by)? Which
# code reads more like the sentence "median mass by species"?
## Part E: STRETCH -----------------------------------------
# E1. filter() ANDs conditions together with commas. Combine three
# conditions to isolate a specific subgroup, then verify by
# counting rows each way.
# E2. Chain filter + mutate + group_by + summarise into one pipe
# that answers a question of your own about the penguins.