ggplot2 Fundamentals

🎨 ggplot2 Fundamentals: Painting Pictures with Data

Imagine you’re an artist with a magical paintbrush. Instead of colors, you use data. Instead of a canvas, you use a screen. Welcome to ggplot2—the artist’s toolkit for data!

🖼️ The Big Picture: What is ggplot2?

Think of ggplot2 like building with LEGO blocks. Each block is a layer. You stack them one by one to create something amazing!

The Magic Formula:

ggplot(data) +
  aesthetics +
  shapes +
  style

It’s like making a sandwich:

• 🍞 Data = the bread (foundation)
• 🥬 Aesthetics = what goes where (lettuce here, tomato there)
• 🧀 Geoms = the shapes (circles, lines, bars)
• 🎨 Themes = how it looks (fancy or simple)

📍 ggplot Basics: Starting Your Masterpiece

Every great painting starts with a blank canvas. In ggplot2, we create our canvas like this:

library(ggplot2)

ggplot(data = mydata)

That’s it! You just made a blank canvas. Nothing shows up yet because we haven’t told it WHAT to draw or HOW to draw it.

Real Example:

# We have car data
ggplot(data = mtcars)

Think of this like saying: “I want to paint something about cars.” But you haven’t picked up your brush yet!

The Plus Sign (+) is Your Best Friend

In ggplot2, we ADD layers with +. It’s like stacking pancakes!

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

Each + adds another layer to your picture.

🎯 Aesthetic Mappings: Telling Data Where to Go

Aesthetics answer the question: “Which data goes where?”

Imagine you’re sorting toys into boxes:

• Red toys → Red box
• Big toys → Big shelf
• Round toys → Round basket

That’s what aes() does—it sorts your data!

The Basic Aesthetics

Example: Making a Scatter Plot

ggplot(mtcars) +
  aes(x = wt, y = mpg) +
  geom_point()

This says:

• x-axis = car weight
• y-axis = miles per gallon
• Draw = dots

Adding Color Magic

ggplot(mtcars) +
  aes(x = wt,
      y = mpg,
      color = factor(cyl)) +
  geom_point()

Now each dot is colored by how many cylinders the car has. 4-cylinder cars might be red, 6-cylinder blue, 8-cylinder green!

Inside vs. Outside aes()

This is SUPER important:

# INSIDE aes() = varies by data
aes(color = species)

# OUTSIDE aes() = same for all
geom_point(color = "blue")

Think of it like:

• Inside aes(): “Color each point based on what team they’re on”
• Outside aes(): “Make ALL points blue, no exceptions”

🔷 Geometric Objects: The Shapes of Your Story

Geoms are the actual shapes you see. They’re like different stamp shapes:

graph TD
    A["geom_point"] --> B["Dots/Scatter"]
    C["geom_line"] --> D["Connected Lines"]
    E["geom_bar"] --> F["Bar Charts"]
    G["geom_histogram"] --> H["Distribution Bars"]
    I["geom_boxplot"] --> J["Box & Whiskers"]

The Most Popular Geoms

1. Points (Scatter plots)

ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point()

Use when: Comparing two numbers

2. Lines

ggplot(economics, aes(x = date, y = pop)) +
  geom_line()

Use when: Showing change over time

3. Bars

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

Use when: Counting categories

4. Histogram

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

Use when: Seeing distribution of one number

5. Boxplot

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

Use when: Comparing distributions across groups

Stacking Multiple Geoms

The coolest part? You can stack geoms like LEGO!

ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point() +
  geom_smooth()

This draws dots AND adds a smooth trend line. Two layers, one picture!

📊 Faceting: Small Multiples Magic

Faceting is like having multiple small TVs instead of one big one. Each TV shows a different channel!

facet_wrap() — Wrapping Around

ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point() +
  facet_wrap(~cyl)

This creates ONE plot for each cylinder value:

• One plot for 4-cylinder cars
• One plot for 6-cylinder cars
• One plot for 8-cylinder cars

The ~ means “split by”—like saying “one plot PER cylinder type.”

facet_grid() — Making a Grid

ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point() +
  facet_grid(am ~ cyl)

This makes a GRID:

• Rows = automatic vs manual transmission
• Columns = 4, 6, or 8 cylinders

graph TD
    subgraph Facet Grid
        A["Row: Manual"]
        B["Row: Automatic"]
        C["Col: 4cyl"]
        D["Col: 6cyl"]
        E["Col: 8cyl"]
    end

Controlling Scales

# Same scale for all (default)
facet_wrap(~cyl)

# Free scales (each plot has its own)
facet_wrap(~cyl, scales = "free")

🎨 ggplot Themes: Dressing Up Your Plot

Themes are like outfits for your plot. The data stays the same, but the LOOK changes!

Built-in Themes

# Default gray
ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point() +
  theme_gray()

# Clean and minimal
theme_minimal()

# Classic look (white background)
theme_classic()

# Black and white
theme_bw()

# Dark theme
theme_dark()

Customizing with theme()

Want to change specific things? Use theme():

ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point() +
  theme(
    plot.title = element_text(size = 20),
    axis.text = element_text(color = "blue"),
    panel.background = element_rect(fill = "white")
  )

Common Theme Elements

Quick Legend Tricks

# Move legend to bottom
theme(legend.position = "bottom")

# Remove legend completely
theme(legend.position = "none")

# Put legend inside the plot
theme(legend.position = c(0.9, 0.9))

📐 ggplot Coordinates: Controlling the View

Coordinates control HOW you see your plot—like adjusting a camera!

Zooming In

ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point() +
  coord_cartesian(xlim = c(2, 4), ylim = c(15, 25))

This zooms into a specific area WITHOUT removing data.

Flipping Axes

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

Turns vertical bars into horizontal bars!

graph LR
    A["Vertical Bar Chart"] -->|coord_flip| B["Horizontal Bar Chart"]

Fixed Aspect Ratio

ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point() +
  coord_fixed(ratio = 1)

Makes sure 1 unit on x = 1 unit on y. Great for maps or when proportions matter!

Polar Coordinates (Pie Charts!)

ggplot(mtcars, aes(x = "", fill = factor(cyl))) +
  geom_bar() +
  coord_polar("y")

This bends a bar chart into a circle—a pie chart!

xlim and ylim Differences

Two ways to set axis limits:

# Method 1: coord_cartesian (KEEPS all data)
coord_cartesian(xlim = c(10, 30))

# Method 2: scale limits (REMOVES data outside)
scale_x_continuous(limits = c(10, 30))

Use coord_cartesian() when you want to ZOOM. Use scale limits when you want to FILTER.

🚀 Putting It All Together

Here’s a complete example using EVERYTHING:

library(ggplot2)

ggplot(data = mtcars) +
  aes(x = wt,
      y = mpg,
      color = factor(cyl),
      size = hp) +
  geom_point(alpha = 0.7) +
  facet_wrap(~am, labeller = labeller(
    am = c("0" = "Automatic", "1" = "Manual")
  )) +
  theme_minimal() +
  theme(legend.position = "bottom") +
  coord_cartesian(ylim = c(10, 35)) +
  labs(
    title = "Car Performance",
    x = "Weight (1000 lbs)",
    y = "Miles Per Gallon",
    color = "Cylinders",
    size = "Horsepower"
  )

What this does:

1. 📊 Data: Uses mtcars dataset
2. 🎯 Aesthetics: Weight on x, MPG on y, color by cylinders, size by horsepower
3. 🔷 Geom: Draws semi-transparent points
4. 📐 Facet: Splits by transmission type
5. 🎨 Theme: Minimal style with bottom legend
6. 📍 Coordinates: Zooms to MPG 10-35
7. 🏷️ Labels: Nice titles for everything

🎓 Key Takeaways

1. Start simple: ggplot(data) + aes() + geom_*()
2. Layer by layer: Add complexity with +
3. Aesthetics inside aes(): For data-driven properties
4. Choose the right geom: Points, lines, bars—match your question
5. Facet for comparison: Small multiples reveal patterns
6. Theme for polish: Make it beautiful
7. Coordinates for control: Zoom, flip, transform

🌟 Remember: ggplot2 is like a language. The more you practice, the more fluent you become. Start with simple plots, then add layers one at a time. Before you know it, you’ll be creating data art!

The Grammar of Graphics in One Sentence:

Take some DATA, map it to AESTHETICS, draw it with GEOMS, maybe FACET it, style it with a THEME, and adjust the COORDINATES.

That’s it. You’re now a ggplot2 artist! 🎨