5  Session 2: R Fundamentals

Master data transformation and visualisation

In Session 1, you ran pre-written code. Today you learn to write your own – to ask questions of data and get visual answers. By the end, you’ll build an indicator chart from scratch.

5.1 Learning Outcomes

  • Understand R’s key data structure (tibble/data frame)
  • Use the pipe operator (%>% or |>) to chain operations
  • Apply core dplyr verbs: filter, select, mutate, summarise, group_by
  • Create multi-layered ggplot2 visualisations
  • Debug common R errors

5.2 Session Structure

5.2.1 Part 1: Homework Review & Setup (30 min)

Homework Review (25 min):

  • Show-and-tell: Volunteers demonstrate their homework modifications
  • Address common questions from homework prep
  • Quick poll: What was most confusing? (Adjust session focus accordingly)

Setting Up Your Practise Script (5 min):

Today you’ll write R code from scratch. You need somewhere to work:

  1. In Positron, open the indicators project if it isn’t already open
  2. File > New File > R Script (or Ctrl+Shift+N)
  3. Save it straight away as scripts/R/session-02-practice.R
  4. Add a header comment at the top:
# Session 2 Practice - R Fundamentals
# Your name, today's date

You’ll use this file for all the “Try It” exercises today. Run individual lines with Ctrl+Enter or highlight a block and run it with Ctrl+Shift+Enter.

5.2.2 Part 2: R Basics - Data Structures (45 min)

Your indicator data arrives as a table of rows and columns – understanding tibbles is how R represents this.

The Tibble: R’s Spreadsheet

library(tidyverse)

# Load a real WECA dataset
area_data <- read_csv(here::here("data", "examples", "area_employment.csv"))

# View it
area_data

# The $ operator accesses a single column (base R syntax)
area_data$employment_rate

Key Concepts:

  • Tibbles are like Excel tables with named columns
  • <- means “assign to” (store a value)
  • $ accesses a column
  • R is case-sensitive: Population ≠ population
  • Comments start with #

BREAK (15 min)

5.2.3 Part 3: The Tidyverse Way - Verbs and Pipes (60 min)

Building an indicator means filtering to your area, calculating rates, and summarising trends – these five verbs do exactly that.

The Pipe Operator:

# Old way (nested, hard to read)
summarise(filter(area_data, population > 100000),
          mean_pop = mean(population))

# Tidyverse way (sequential, readable)
area_data %>%
  filter(population > 100000) %>%
  summarise(mean_pop = mean(population))

Try It (5 min): Rewrite this nested expression using pipes: select(filter(area_data, employment_rate > 0.5), area, employment_rate)

The Five Core Verbs:

  1. filter() - Keep rows that match a condition
  2. select() - Keep specific columns
  3. mutate() - Create new columns
  4. summarise() - Calculate summary statistics
  5. group_by() - Do calculations by category

Try It (25 min): Using your example dataset, work through each verb:

  1. filter() – keep only rows for a specific area
  2. select() – keep just the area, year, and value columns
  3. mutate() – add a new column calculating a rate or percentage
  4. summarise() – calculate the mean and max of a numeric column
  5. group_by() %>% summarise() – calculate summary statistics by area

Pair up for these exercises. Person A writes the filter() and select() steps; Person B writes mutate() and summarise(). Swap roles for the ggplot2 exercise in Part 4.

BREAK (15 min)

5.2.4 Part 4: ggplot2 - Grammar of Graphics (60 min)

Every indicator chapter needs at least one chart. This is how you build them.

# Layer 1: Data + aesthetic mapping
ggplot(bus_data, aes(x = year, y = ridership))

# Layer 2: Add geometry
ggplot(bus_data, aes(x = year, y = ridership)) +
  geom_line()

# Layer 3: Add styling
ggplot(bus_data, aes(x = year, y = ridership)) +
  geom_line(colour = get_weca_color("forest_green"), linewidth = 1) +
  geom_point(size = 2) +
  labs(title = "Bus Ridership Over Time",
       x = "Year",
       y = "Ridership (millions)") +
  theme_weca()

Key ggplot2 Geometries:

  • geom_line() - Line charts (trends over time)
  • geom_point() - Scatter plots (relationships)
  • geom_col() - Bar charts (comparisons)
  • geom_smooth() - Trend lines

Try It (10 min): Create a bar chart of employment rate by area using geom_col(). Add WECA colours and labels with labs().

5.2.5 Part 5: Debugging R Errors (30 min)

Common Error Messages and Fixes:

  1. could not find function "filter" → Load library: library(tidyverse)
  2. object 'data' not found → Run the chunk that loads the data first
  3. unexpected symbol → Cheque for missing commas, quotes, or parentheses
  4. 'x' and 'y' lengths differ → Data columns have different numbers of rows

Debugging Strategy:

  1. Read the error message (bottom-up in stack trace)
  2. Cheque the line number mentioned
  3. Look for typos in variable/column names
  4. Run code chunk-by-chunk to isolate the problem
  5. Use View(data) to inspect your data frame, use the data explorer or use data |> glimpse()

5.2.6 Part 6: Wrap-up & Homework (15 min)

Reflection (5 min):

Before we finish, take 2 minutes to write down:

  1. One thing you understand now that you didn’t at the start
  2. One thing that’s still unclear or you’d like more practise with

Which dplyr verb felt most natural? Which was most confusing? Share with the group if you’re comfortable.

Homework (1-2 hours):

  1. Complete: The “R Basics” chapter in R for Data Science
  2. Practise: Using your assigned example dataset:
    • Calculate 3 summary statistics (min, max, mean)
    • Create 2 different chart types
    • Add WECA theme and appropriate labels
  3. Create: A new indicator section that includes data loading, transformation, visualisation, and findings paragraph
  4. Prepare: Bring a real WECA dataset (CSV) for Session 3