Chapter 26 Introduction to ggplot2

What You’ll Learn:

  • Grammar of Graphics principles
  • Basic ggplot2 structure
  • Common plotting errors
  • Aesthetics and geoms
  • Layer system

Key Errors Covered: 20+ ggplot2 errors

Difficulty: ⭐⭐ Intermediate to ⭐⭐⭐ Advanced

26.1 Introduction

ggplot2 revolutionized R graphics with the Grammar of Graphics:

library(ggplot2)

# Basic plot
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point()

But ggplot2 has unique error patterns. Let’s master them!

26.2 ggplot2 Structure

💡 Key Insight: Grammar of Graphics

# Three essential components:
# 1. Data
# 2. Aesthetic mappings (aes)
# 3. Geometric objects (geom)

# Basic structure
ggplot(data = mtcars, mapping = aes(x = mpg, y = hp)) +
  geom_point()

# Shortened (common)
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point()

# Can specify aes in geom instead
ggplot(mtcars) +
  geom_point(aes(x = mpg, y = hp))

# Or mix (useful for multiple layers)
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  geom_smooth()
#> `geom_smooth()` using method = 'loess' and formula = 'y ~ x'

Key principle: Build plots in layers with +

26.3 Error #1: object not found in aes()

⭐ BEGINNER 🔍 SCOPE

26.3.1 The Error

ggplot(mtcars, aes(x = mpg, y = horsepower)) +
  geom_point()
#> Error in `geom_point()`:
#> ! Problem while computing aesthetics.
#> ℹ Error occurred in the 1st layer.
#> Caused by error:
#> ! object 'horsepower' not found

🔴 ERROR

Error in FUN(X[[i]], ...) : object 'horsepower' not found

26.3.2 What It Means

The column name doesn’t exist in the data.

26.3.3 Common Causes

# Typo in column name
ggplot(mtcars, aes(x = mpgg, y = hp)) +
  geom_point()
#> Error in `geom_point()`:
#> ! Problem while computing aesthetics.
#> ℹ Error occurred in the 1st layer.
#> Caused by error:
#> ! object 'mpgg' not found

# Wrong dataset
ggplot(iris, aes(x = mpg, y = hp)) +
  geom_point()
#> Error in `geom_point()`:
#> ! Problem while computing aesthetics.
#> ℹ Error occurred in the 1st layer.
#> Caused by error:
#> ! object 'hp' not found

# Forgot to create column
ggplot(mtcars, aes(x = mpg, y = efficiency)) +
  geom_point()
#> Error in `geom_point()`:
#> ! Problem while computing aesthetics.
#> ℹ Error occurred in the 1st layer.
#> Caused by error:
#> ! object 'efficiency' not found

26.3.4 Solutions

✅ SOLUTION 1: Check Column Names

# Verify columns
names(mtcars)
#>  [1] "mpg"        "cyl"        "disp"       "hp"         "drat"      
#>  [6] "wt"         "qsec"       "vs"         "am"         "gear"      
#> [11] "carb"       "cyl_factor"

# Use correct name
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point()

✅ SOLUTION 2: Create Column First

library(dplyr)

mtcars %>%
  mutate(efficiency = mpg / hp) %>%
  ggplot(aes(x = hp, y = efficiency)) +
  geom_point()

26.4 Error #2: Using + vs %>%

⭐ BEGINNER 🔤 SYNTAX

26.4.1 The Error

library(dplyr)

mtcars %>%
  filter(cyl == 4) %>%
  ggplot(aes(x = mpg, y = hp)) %>%  # Wrong operator!
  geom_point()
#> Error in `geom_point()`:
#> ! `mapping` must be created by `aes()`.
#> ✖ You've supplied a <ggplot2::ggplot> object.
#> ℹ Did you use `%>%` or `|>` instead of `+`?

🔴 ERROR

Error in geom_point(.) : 
  Cannot use `+` with a ggplot object. Did you accidentally use `%>%` instead of `+`?

26.4.2 What It Means

Must use + to add ggplot2 layers, not %>%.

26.4.3 Solutions

✅ SOLUTION: Use + for ggplot layers

# Correct: + for ggplot
mtcars %>%
  filter(cyl == 4) %>%
  ggplot(aes(x = mpg, y = hp)) +  # Use +
  geom_point()

# Pipe data into ggplot, then use +
mtcars %>%
  filter(cyl == 4) %>%
  ggplot(aes(x = mpg, y = hp)) +
  geom_point() +
  theme_minimal()

26.5 Aesthetics (aes)

💡 Key Insight: Aesthetic Mappings

# Map variables to visual properties
ggplot(mtcars, aes(x = mpg, y = hp, color = factor(cyl))) +
  geom_point()

# Common aesthetics:
# x, y - position
# color - point/line color
# fill - area fill color
# size - point/line size
# shape - point shape
# alpha - transparency
# linetype - line pattern

# Multiple aesthetics
ggplot(mtcars, aes(x = mpg, y = hp, 
                   color = factor(cyl),
                   size = wt)) +
  geom_point()

# Set vs map
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point(color = "blue")  # Set: all points blue

ggplot(mtcars, aes(x = mpg, y = hp, color = factor(cyl))) +
  geom_point()  # Map: color varies by cyl

26.6 Error #3: Aesthetic outside aes()

⭐⭐ INTERMEDIATE 🧠 LOGIC

26.6.1 The Error

# Trying to map cyl to color outside aes()
ggplot(mtcars) +
  geom_point(aes(x = mpg, y = hp), color = cyl)
#> Error: object 'cyl' not found

🔴 ERROR

Error in layer(...) : object 'cyl' not found

26.6.2 What It Means

Variable mappings must be inside aes().

26.6.3 Common Causes

# Want color by variable
ggplot(mtcars) +
  geom_point(aes(x = mpg, y = hp), color = factor(cyl))
#> Error: object 'cyl' not found

# Want size by variable
ggplot(mtcars) +
  geom_point(aes(x = mpg, y = hp), size = wt)
#> Error: object 'wt' not found

26.6.4 Solutions

✅ SOLUTION: Put Variable Mappings in aes()

# Correct: color mapping inside aes
ggplot(mtcars) +
  geom_point(aes(x = mpg, y = hp, color = factor(cyl)))

# Can be in ggplot() aes
ggplot(mtcars, aes(x = mpg, y = hp, color = factor(cyl))) +
  geom_point()

# Fixed values go OUTSIDE aes
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point(color = "blue", size = 3)  # All points same

⚠️ Common Confusion: Inside vs Outside aes()

# INSIDE aes(): varies by data
ggplot(mtcars, aes(x = mpg, y = hp, color = factor(cyl))) +
  geom_point()  # Color varies by cyl

# OUTSIDE aes(): fixed for all
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point(color = "red")  # All points red

# Wrong: puts string in aes
ggplot(mtcars, aes(x = mpg, y = hp, color = "red")) +
  geom_point()  # Creates legend for "red"!

26.7 Common geoms

💡 Key Insight: Geometric Objects

# Points
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point()

# Lines
ggplot(economics, aes(x = date, y = unemploy)) +
  geom_line()

# Bars
ggplot(mtcars, aes(x = factor(cyl))) +
  geom_bar()

# Histogram
ggplot(mtcars, aes(x = mpg)) +
  geom_histogram(bins = 10)

# Boxplot
ggplot(mtcars, aes(x = factor(cyl), y = mpg)) +
  geom_boxplot()

# Smooth
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  geom_smooth(method = "lm")
#> `geom_smooth()` using formula = 'y ~ x'

# Text
ggplot(mtcars, aes(x = mpg, y = hp, label = rownames(mtcars))) +
  geom_text(size = 3)

26.8 Error #4: stat_count() requires x or y

⭐ BEGINNER 📋 ARGS

26.8.1 The Error

ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_bar()
#> Error in `geom_bar()`:
#> ! Problem while computing stat.
#> ℹ Error occurred in the 1st layer.
#> Caused by error in `setup_params()`:
#> ! `stat_count()` must only have an x or y aesthetic.

🔴 ERROR

Error in `geom_bar()`:
! Problem while computing stat.
ℹ Error occurred in the 1st layer.
Caused by error:
! `stat_count()` must only have an `x` or `y` aesthetic.

26.8.2 What It Means

geom_bar() is for counts. For pre-computed heights, use geom_col().

26.8.3 Solutions

✅ SOLUTION 1: Use geom_col() for Heights

# Pre-computed values
data <- data.frame(
  category = c("A", "B", "C"),
  value = c(10, 15, 20)
)

ggplot(data, aes(x = category, y = value)) +
  geom_col()

# Or use stat = "identity" with geom_bar
ggplot(data, aes(x = category, y = value)) +
  geom_bar(stat = "identity")

✅ SOLUTION 2: Use geom_bar() for Counts

# Count occurrences
ggplot(mtcars, aes(x = factor(cyl))) +
  geom_bar()

# Equivalent to
mtcars %>%
  count(cyl) %>%
  ggplot(aes(x = factor(cyl), y = n)) +
  geom_col()

26.9 Faceting

💡 Key Insight: Small Multiples

# Facet by one variable
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  facet_wrap(~ cyl)

# Facet by two variables
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  facet_grid(cyl ~ gear)

# Free scales
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  facet_wrap(~ cyl, scales = "free")

# Number of columns
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  facet_wrap(~ cyl, ncol = 2)

26.10 Themes and Customization

🎯 Best Practice: Customizing Plots

# Built-in themes
p <- ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point()

p + theme_minimal()
p + theme_classic()
p + theme_bw()

# Custom labels
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  labs(
    title = "Fuel Efficiency vs Horsepower",
    subtitle = "Motor Trend Car Road Tests",
    x = "Miles per Gallon",
    y = "Horsepower",
    caption = "Source: mtcars dataset"
  )

# Customize theme elements
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  theme_minimal() +
  theme(
    plot.title = element_text(face = "bold", size = 14),
    axis.text = element_text(size = 10),
    panel.grid.minor = element_blank()
  )

26.11 Error #5: Non-numeric variable for histogram

⭐ BEGINNER 🔢 TYPE

26.11.1 The Error

ggplot(mtcars, aes(x = factor(cyl))) +
  geom_histogram()
#> Error in `geom_histogram()`:
#> ! Problem while computing stat.
#> ℹ Error occurred in the 1st layer.
#> Caused by error in `setup_params()`:
#> ! `stat_bin()` requires a continuous x aesthetic.
#> ✖ the x aesthetic is discrete.
#> ℹ Perhaps you want `stat="count"`?

🔴 ERROR

Error in `geom_histogram()`:
! `stat_bin()` requires a numeric `x` variable

26.11.2 What It Means

Histograms need continuous numeric data, not categorical.

26.11.3 Solutions

✅ SOLUTION: Use Appropriate Geom

# For categorical: use geom_bar
ggplot(mtcars, aes(x = factor(cyl))) +
  geom_bar()

# For continuous: histogram works
ggplot(mtcars, aes(x = mpg)) +
  geom_histogram(bins = 10)

# Density plot alternative
ggplot(mtcars, aes(x = mpg)) +
  geom_density()

26.12 Saving Plots

🎯 Best Practice: Save Plots

# Create plot
p <- ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  theme_minimal()

# Save with ggsave
ggsave("plot.png", p, width = 6, height = 4, dpi = 300)

# Or save last plot
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point()

ggsave("last_plot.png", width = 6, height = 4)

# Different formats
ggsave("plot.pdf", p)
ggsave("plot.svg", p)
ggsave("plot.jpg", p)

26.13 Common Patterns

🎯 Best Practice: Common Plot Types

# Scatterplot with trend line
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point(aes(color = factor(cyl))) +
  geom_smooth(method = "lm", se = FALSE) +
  theme_minimal()
#> `geom_smooth()` using formula = 'y ~ x'

# Grouped bar chart
mtcars %>%
  count(cyl, gear) %>%
  ggplot(aes(x = factor(cyl), y = n, fill = factor(gear))) +
  geom_col(position = "dodge")

# Boxplot with points
ggplot(mtcars, aes(x = factor(cyl), y = mpg)) +
  geom_boxplot() +
  geom_jitter(width = 0.2, alpha = 0.3)

# Time series
ggplot(economics, aes(x = date, y = unemploy)) +
  geom_line() +
  theme_minimal() +
  labs(title = "US Unemployment Over Time")

26.14 Summary

Key Takeaways:

  1. Three components - Data, aes(), geom
  2. Use + not %>% - Add layers with +
  3. Variables in aes() - Fixed values outside
  4. geom_bar() vs geom_col() - Counts vs heights
  5. Check column names - Before plotting
  6. Histograms need numeric - Use geom_bar() for categorical
  7. Build in layers - Add components step by step

Quick Reference:

Error Cause Fix
object not found Column doesn’t exist Check names(data)
Can’t use + Used %>% instead of + Use + for ggplot layers
object ‘var’ not found Variable outside aes Put in aes()
stat_count requires x or y geom_bar with y Use geom_col()
requires numeric x Non-numeric histogram Use appropriate geom

Basic Structure:

# Template
ggplot(data, aes(x = var1, y = var2)) +
  geom_point() +
  theme_minimal()

# With pipes
data %>%
  filter(condition) %>%
  ggplot(aes(x = var1, y = var2)) +  # Use +
  geom_point() +
  labs(title = "Plot Title")

# Common aesthetics
aes(
  x = var,           # x-axis
  y = var,           # y-axis
  color = var,       # point/line color
  fill = var,        # area fill
  size = var,        # size
  shape = var,       # point shape
  alpha = var,       # transparency
  linetype = var     # line pattern
)

# Common geoms
geom_point()         # scatter
geom_line()          # line
geom_bar()           # bar (counts)
geom_col()           # bar (heights)
geom_histogram()     # histogram
geom_boxplot()       # boxplot
geom_smooth()        # trend line

Best Practices:

# ✅ Good
ggplot(data, aes(x = var1, y = var2, color = var3)) +
  geom_point() +
  theme_minimal()

data %>% filter(...) %>%
  ggplot(aes(x = var)) +  # + not %>%
  geom_histogram()

# ❌ Avoid
ggplot(data, aes(x = var1, y = var2)) %>%  # Wrong operator
  geom_point()

ggplot(data) +
  geom_point(aes(x = var), color = other_var)  # Should be in aes

geom_histogram(aes(x = factor_var))  # Need numeric

26.15 Exercises

📝 Exercise 1: Basic Plot

Create a scatterplot of mtcars: 1. mpg vs hp 2. Color by cyl 3. Size by wt 4. Add title and labels 5. Use theme_minimal()

📝 Exercise 2: Error Fixing

Fix these errors:

# Error 1
ggplot(mtcars, aes(x = mpg, y = hp)) %>%
  geom_point()

# Error 2
ggplot(mtcars) +
  geom_point(aes(x = mpg, y = hp), color = cyl)

# Error 3
ggplot(mtcars, aes(x = factor(cyl), y = mpg)) +
  geom_histogram()

📝 Exercise 3: Multiple Geoms

Create a plot with: 1. Points for raw data 2. Smooth line for trend 3. Faceted by cyl 4. Custom colors

📝 Exercise 4: Bar Chart

Using mtcars: 1. Count cars by cyl 2. Fill by gear 3. Dodge position 4. Add labels

26.16 Exercise Answers

Click to see answers

Exercise 1:

ggplot(mtcars, aes(x = mpg, y = hp, color = factor(cyl), size = wt)) +
  geom_point(alpha = 0.7) +
  labs(
    title = "Fuel Efficiency vs Horsepower",
    subtitle = "Motor Trend Car Road Tests",
    x = "Miles per Gallon",
    y = "Horsepower",
    color = "Cylinders",
    size = "Weight (1000 lbs)"
  ) +
  theme_minimal() +
  theme(
    plot.title = element_text(face = "bold", size = 14),
    legend.position = "right"
  )

Exercise 2:

# Error 1: Use + not %>%
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point()

# Error 2: Put variable in aes()
ggplot(mtcars) +
  geom_point(aes(x = mpg, y = hp, color = factor(cyl)))

# Error 3: Use geom_boxplot for this
ggplot(mtcars, aes(x = factor(cyl), y = mpg)) +
  geom_boxplot()

# Or if want histogram of mpg
ggplot(mtcars, aes(x = mpg)) +
  geom_histogram(bins = 10)

Exercise 3:

ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point(aes(color = factor(cyl)), size = 3, alpha = 0.6) +
  geom_smooth(method = "lm", se = FALSE, color = "black", linetype = "dashed") +
  facet_wrap(~ cyl, labeller = label_both) +
  scale_color_manual(values = c("4" = "#E41A1C", "6" = "#377EB8", "8" = "#4DAF4A")) +
  labs(
    title = "MPG vs HP by Number of Cylinders",
    x = "Miles per Gallon",
    y = "Horsepower",
    color = "Cylinders"
  ) +
  theme_bw()
#> `geom_smooth()` using formula = 'y ~ x'

Exercise 4:

library(dplyr)

mtcars %>%
  count(cyl, gear) %>%
  ggplot(aes(x = factor(cyl), y = n, fill = factor(gear))) +
  geom_col(position = "dodge") +
  labs(
    title = "Car Count by Cylinders and Gears",
    x = "Number of Cylinders",
    y = "Count",
    fill = "Number of Gears"
  ) +
  theme_minimal() +
  scale_fill_brewer(palette = "Set2")