Charts: compare & relate

Set up

library(tidyverse)
students <- read_csv("data/students.csv", show_col_types = FALSE) %>%
  mutate(Motivation = factor(Motivation, levels = c("Low", "Medium", "High")))

Boxplots: compare groups

R codeCopy

ggplot(students, aes(x = Motivation, y = ExamScore)) +
  geom_boxplot()

How to read one box:

Add geom_jitter() to show every student as a dot:

R codeCopy

ggplot(students, aes(x = Motivation, y = ExamScore)) +
  geom_boxplot() +
  geom_jitter(width = 0.15, alpha = 0.4)

alpha sets see-through-ness (0 to 1). The jitter is random, so your dots will differ a little.

Scatter plots: two numbers

R codeCopy

ggplot(students, aes(x = StudyHours, y = ExamScore)) +
  geom_point(alpha = 0.6)

Output

Warning message:
Removed 6 rows containing missing values or values outside the scale
range (`geom_point()`).

More study, higher scores. The warning: 6 students have no study hours, so they can't be drawn.

Add a straight trend line with geom_smooth(method = "lm"):

R codeCopy

ggplot(students, aes(x = StudyHours, y = ExamScore)) +
  geom_point(alpha = 0.6) +
  geom_smooth(method = "lm")

Output

`geom_smooth()` using formula = 'y ~ x'
Warning messages:
1: Removed 6 rows containing non-finite outside the scale range
(`stat_smooth()`).
2: Removed 6 rows containing missing values or values outside the scale
range (`geom_point()`).

The grey band shows how sure we are about the line.

Make it report-ready

R codeCopy

ggplot(students, aes(x = StudyHours, y = ExamScore)) +
  geom_point(alpha = 0.6, colour = "steelblue") +
  geom_smooth(method = "lm", colour = "black") +
  labs(title = "Study time and exam score",
       x = "Study hours per week",
       y = "Exam score (out of 100)") +
  theme_classic()

Output

`geom_smooth()` using formula = 'y ~ x'
Warning messages:
1: Removed 6 rows containing non-finite outside the scale range
(`stat_smooth()`).
2: Removed 6 rows containing missing values or values outside the scale
range (`geom_point()`).

• labs() sets the title and axis labels.
• theme_classic() gives a clean look. Try theme_minimal() too.

Save the last chart as an image file in your project folder:

ggsave("study-exam.png", width = 6, height = 4, dpi = 300)

Which chart?

Exercises

Exercise 9.1 Dataset: students.csv

Boxplot

Compare Stress between students living in a hall and at home, with a boxplot and jittered dots.

Show answerHide answer

R codeCopy

ggplot(students, aes(x = Living, y = Stress)) +
  geom_boxplot() +
  geom_jitter(width = 0.15, alpha = 0.3)

The two groups look very similar.

Exercise 9.2 Dataset: students.csv

Scatter plot

Plot Sleep (x) against Stress (y) with a trend line, clear labels and theme_minimal(). What does the line show?

Show answerHide answer

R codeCopy

ggplot(students, aes(x = Sleep, y = Stress)) +
  geom_jitter(width = 0, height = 0.2, alpha = 0.5) +
  geom_smooth(method = "lm") +
  labs(x = "Hours of sleep per night", y = "Stress (1 to 10)") +
  theme_minimal()

Output

`geom_smooth()` using formula = 'y ~ x'
Warning messages:
1: Removed 10 rows containing non-finite outside the scale range
(`stat_smooth()`).
2: Removed 10 rows containing missing values or values outside the scale
range (`geom_point()`).

The line goes down: students who sleep more tend to feel less stressed. geom_jitter(height = 0.2) nudges the points up and down a little, because stress only takes whole numbers and many points would sit on top of each other.

Key points

• Boxplot: a number across groups. Scatter plot: two numbers.
• geom_smooth(method = "lm") adds a straight trend line.
• labs() for labels, theme_classic() for a clean look, ggsave() to save.