r‑statistics.co by Selva Prabhakaran

About

ggplot2 Cheat Sheet: Quick Reference for Geoms, Scales, Themes and More

This cheat sheet is your one-page reference for ggplot2 — covering the layer grammar, every major geom, aesthetic mappings, scale functions, facets, and theme customization, each with a minimal working example you can run and modify.

Introduction

ggplot2 follows a layered grammar: start with data and a coordinate system, add geometric layers (geoms), control how data maps to visual properties (aesthetics and scales), and polish with themes. Every chart in ggplot2 is built by composing these layers using +.

The basic template is always:

library(ggplot2) # The universal ggplot2 template # ggplot(data, aes(x = ..., y = ...)) + geom_*() + scale_*() + theme_*() # Verify it works with a minimal example p_base <- ggplot(mtcars, aes(x = wt, y = mpg)) + geom_point() + labs(title = "Weight vs MPG", x = "Weight (1000 lbs)", y = "Miles per Gallon") p_base


  


  






Each section below shows the key functions in that category with quick examples. All code shares a single session — mtcars and mpg are always available.



What Are the Essential Geoms in ggplot2?



A geom defines the visual mark drawn for each data point. Add multiple geoms to the same chart by layering them with +.







Geom
Use For
Key Arguments






geom_point()
Scatter plots
size, shape, alpha




geom_line()
Line charts (sorted by x)
linewidth, linetype




geom_bar()
Count frequency from raw data
fill, position




geom_col()
Bar heights from pre-computed values
fill, width




geom_histogram()
Distribution of one continuous variable
binwidth, bins, fill




geom_density()
Smooth density curve
adjust, fill, alpha




geom_boxplot()
Median, quartiles, outliers
fill, notch




geom_violin()
Full distribution shape per group
fill, draw_quantiles




geom_smooth()
Trend / regression line
method, se




geom_text()
Text labels at data positions
label, hjust, vjust




geom_hline()
Horizontal reference line
yintercept, linetype




geom_vline()
Vertical reference line
xintercept, linetype




geom_abline()
Line by slope and intercept
slope, intercept




geom_area()
Filled area under a line
fill, alpha




geom_tile()
Heatmap tiles
fill




geom_bin2d()
2D density bins
bins, fill




geom_jitter()
Scatter with random noise (reduces overlap)
width, height




geom_errorbar()
Error bars
ymin, ymax, width




geom_ribbon()
Shaded band between y bounds
ymin, ymax, fill




geom_segment()
Arbitrary line segments
x, xend, y, yend









  


    
# Quick geom gallery - 4 charts on one dataset
p_point <- ggplot(mpg, aes(x = displ, y = hwy)) +
  geom_point(color = "steelblue", alpha = 0.6) +
  labs(title = "geom_point()", x = "Displacement", y = "MPG")

p_smooth <- ggplot(mpg, aes(x = displ, y = hwy)) +
  geom_point(alpha = 0.3) +
  geom_smooth(method = "lm", color = "firebrick") +
  labs(title = "geom_point() + geom_smooth()", x = "Displacement", y = "MPG")

p_hist <- ggplot(mpg, aes(x = hwy)) +
  geom_histogram(binwidth = 3, fill = "steelblue", color = "white") +
  labs(title = "geom_histogram()", x = "Highway MPG")

p_box <- ggplot(mpg, aes(x = drv, y = hwy, fill = drv)) +
  geom_boxplot(show.legend = FALSE) +
  scale_fill_brewer(palette = "Set2") +
  labs(title = "geom_boxplot()", x = "Drive Type", y = "Highway MPG")

# Show all four
library(patchwork)
(p_point | p_smooth) / (p_hist | p_box)


    


      

      

    


    


  


  






TIP: Stack multiple geoms to build richer charts. A common pattern: geom_boxplot() + geom_jitter() shows both the summary statistics and the raw data distribution simultaneously.



Try it: Replace geom_boxplot() with geom_violin() in p_box. How does the shape of the distribution change compared to the boxplot?





  


    
ex_violin <- ggplot(mpg, aes(x = drv, y = hwy, fill = drv)) +
  geom_violin(show.legend = FALSE) +
  scale_fill_brewer(palette = "Set2") +
  labs(x = "Drive Type", y = "Highway MPG")

ex_violin


    


      

      

    


    


  


  






How Do Aesthetics and aes() Work?



Aesthetics map data columns to visual properties. Mappings inside aes() are data-driven. Values set outside aes() are fixed for all data points.







Aesthetic
Controls
Variable Types
Example






x, y
Position
Continuous, discrete, date
aes(x = date, y = value)




color
Point/line/text color
Continuous → gradient; Discrete → palette
aes(color = species)




fill
Interior fill color
Continuous → gradient; Discrete → palette
aes(fill = category)




size
Point size, line thickness
Continuous numeric
aes(size = population)




shape
Point shape (16 = circle, 17 = triangle…)
Discrete (≤ 6 levels)
aes(shape = group)




alpha
Transparency (0 = invisible, 1 = opaque)
Continuous numeric
aes(alpha = confidence)




linetype
Line style
Discrete
aes(linetype = type)




label
Text content
Character
aes(label = name)




group
Grouping without visual change
Discrete
aes(group = id)









  


    
# Demonstrate 4 aesthetics mapped simultaneously
p_aes <- ggplot(mpg, aes(
    x     = displ,
    y     = hwy,
    color = drv,      # categorical → discrete color
    size  = cyl,      # numeric → size scale
    alpha = cty       # numeric → transparency
  )) +
  geom_point() +
  scale_color_brewer(palette = "Set1",
                     labels = c("4"="4WD","f"="Front","r"="Rear")) +
  labs(
    title = "Four aesthetics in one scatter plot",
    x     = "Displacement (L)",
    y     = "Highway MPG",
    color = "Drive",
    size  = "Cylinders",
    alpha = "City MPG"
  )

p_aes


    


      

      

    


    


  


  






KEY INSIGHT: color = "blue" (outside aes()) sets all points to blue. aes(color = drv) maps the drv column to color — one color per unique value. Putting a column name outside aes() (e.g., color = drv) is a common error that maps the literal string, not the column.



Try it: Move size = cyl from inside aes() to outside as size = 3. How does the chart change?





  


    
ex_fixed_size <- ggplot(mpg, aes(x = displ, y = hwy, color = drv)) +
  geom_point(size = 3, alpha = 0.7) +  # fixed size for all points
  scale_color_brewer(palette = "Set1")

ex_fixed_size


    


      

      

    


    


  


  






How Do You Customize Scales for Axes and Colors?



Scale functions control how data values map to visual properties — axis range, breaks, labels, color palettes, size ranges. Every aesthetic has a corresponding scale_*() family.



Axis scales:







Function
Controls
Key Args






scale_x_continuous()
Continuous x-axis
limits, breaks, labels, trans




scale_y_continuous()
Continuous y-axis
limits, breaks, labels, expand




scale_x_discrete()
Categorical x-axis
limits (reorder), labels




scale_x_date()
Date x-axis
date_breaks, date_labels




scale_x_log10()
Log10 x-axis
—







Color/fill scales:







Function
For
When to Use






scale_color_brewer()
Discrete color
ColorBrewer qualitative/sequential




scale_fill_brewer()
Discrete fill
Same, for bar/tile interiors




scale_color_distiller()
Continuous color
ColorBrewer interpolated




scale_fill_distiller()
Continuous fill
Same, for fill




scale_color_viridis_c()
Continuous color
Perceptually uniform, colorblind-safe




scale_color_viridis_d()
Discrete color
Same but for categories




scale_color_manual()
Discrete color
Specify exact hex/named colors




scale_color_gradient()
Continuous, 2-color
Simple low-to-high gradient




scale_color_gradient2()
Continuous, diverging
Low-mid-high with midpoint









  


    
# Demonstrate scale customization
p_scales <- ggplot(mpg, aes(x = displ, y = hwy, color = hwy)) +
  geom_point(size = 3) +
  scale_x_continuous(
    breaks = seq(1, 7, by = 1),
    limits = c(1, 7),
    labels = function(x) paste0(x, "L")
  ) +
  scale_y_continuous(
    breaks = seq(10, 45, by = 5),
    expand = expansion(mult = c(0.05, 0.1))
  ) +
  scale_color_viridis_c(
    option = "plasma",
    name   = "Hwy MPG"
  ) +
  labs(
    title = "Scale customization: axis breaks + labels + color",
    x     = "Engine Displacement",
    y     = "Highway MPG"
  )

p_scales


    


      

      

    


    


  


  






TIP: labels = scales::comma formats numeric axis labels with thousand separators. labels = scales::percent adds a % sign. labels = scales::dollar adds a $ prefix. Load the scales package once and use these as shorthand throughout your project.



Try it: Add scale_x_log10() to any scatter plot. How does a log-transformed axis change the appearance of the data distribution?





  


    
ex_log <- ggplot(diamonds, aes(x = carat, y = price)) +
  geom_point(alpha = 0.05, color = "steelblue") +
  scale_x_log10() +
  scale_y_log10(labels = scales::comma)

ex_log


    


      

      

    


    


  


  






How Do You Use Facets to Create Multi-Panel Charts?



Facets split your data by a variable and draw one panel per group — all with the same axes and geoms, making comparison effortless.







Function
What it Does
Key Args






facet_wrap(~ var)
Wrap panels into rows and columns automatically
nrow, ncol, scales, labeller




facet_wrap(~ var1 + var2)
Facet by two variables, wrapped
Same




facet_grid(rows ~ cols)
Grid layout: one variable per axis
scales, space, labeller




facet_grid(. ~ cols)
Facet columns only
—




facet_grid(rows ~ .)
Facet rows only
—









  


    
# facet_wrap: one panel per drive type
p_facet <- ggplot(mpg, aes(x = displ, y = hwy, color = class)) +
  geom_point(alpha = 0.7) +
  geom_smooth(method = "lm", se = FALSE, color = "grey30",
              linewidth = 0.7, linetype = "dashed") +
  facet_wrap(
    ~ drv,
    nrow     = 1,
    labeller = labeller(drv = c("4"="4WD","f"="Front-wheel","r"="Rear-wheel"))
  ) +
  scale_color_brewer(palette = "Set2") +
  labs(
    title  = "facet_wrap(): one panel per drive type",
    x      = "Displacement (L)",
    y      = "Highway MPG",
    color  = "Class"
  ) +
  theme_minimal()

p_facet


    


      

      

    


    


  


  






The scales argument controls axis behavior across panels:







scales =
Behavior






"fixed" (default)
All panels share the same x and y range




"free_x"
Each panel has its own x-axis range




"free_y"
Each panel has its own y-axis range




"free"
Each panel has independent x and y ranges







KEY INSIGHT: scales = "free_y" is essential when comparing metrics with very different units or magnitudes (e.g., count vs. percentage). Without it, a metric with small values appears as a flat line at the bottom of the panel.



Try it: Add scales = "free" to facet_wrap() in p_facet. How does this change the axes across the three panels?





  


    
ex_free_scales <- ggplot(mpg, aes(x = displ, y = hwy)) +
  geom_point(alpha = 0.6, color = "steelblue") +
  facet_wrap(~ drv, scales = "free")

ex_free_scales


    


      

      

    


    


  


  






How Do You Apply Themes and Polish Your Chart?



Themes control all non-data visual elements: background, grid lines, axis text, legend position, font sizes, and more.



Built-in themes:







Theme
Style






theme_gray()
Default — grey background, white grid




theme_bw()
White background, black grid frame




theme_minimal()
White background, light grey grid, no border




theme_classic()
White background, no grid, clean axes




theme_void()
No axes, no background — for maps and diagrams




theme_dark()
Dark background — good for viridis color scales




theme_light()
Light grey lines, white background







**Customizing with theme():**





  


    
# Build a fully customized theme
p_theme <- ggplot(mpg, aes(x = class, fill = drv)) +
  geom_bar(position = "dodge") +
  scale_fill_brewer(palette = "Set2",
                    labels = c("4"="4WD","f"="Front","r"="Rear")) +
  labs(
    title    = "Custom theme demonstration",
    subtitle = "Every visual element is controllable via theme()",
    x        = "Vehicle Class",
    y        = "Count",
    fill     = "Drive Type",
    caption  = "Source: ggplot2::mpg"
  ) +
  theme_minimal(base_size = 13) +
  theme(
    plot.title       = element_text(face = "bold", size = 16),
    plot.subtitle    = element_text(color = "grey50", size = 11),
    axis.text.x      = element_text(angle = 30, hjust = 1),
    panel.grid.major.x = element_blank(),
    legend.position  = "top",
    legend.title     = element_text(face = "bold"),
    plot.caption     = element_text(color = "grey60", hjust = 1)
  )

p_theme


    


      

      

    


    


  


  






**Key theme() element functions:**







Function
Controls






element_text(face, size, color, angle, hjust, vjust)
Text elements




element_line(color, linewidth, linetype)
Line elements




element_rect(fill, color, linewidth)
Rectangle/border elements




element_blank()
Remove an element entirely







**Common theme() targets:**







Target
What It Affects






plot.title
Chart title




plot.subtitle
Subtitle text




plot.caption
Caption at bottom-right




axis.title.x / y
Axis label text




axis.text.x / y
Axis tick labels




panel.grid.major / minor
Grid lines




panel.background
Plot area background




legend.position
"top", "bottom", "left", "right", "none"




strip.text
Facet panel labels







TIP: Set a theme globally for all charts in a session with theme_set(theme_minimal()). Every subsequent ggplot() call inherits that theme without repeating it. You can still override individual elements with + theme(...) on specific charts.



Try it: Replace theme_minimal() with theme_classic(). How does the chart's overall feel change?





  


    
ex_classic <- p_theme %+% theme_classic()
ex_classic


    


      

      

    


    


  


  






Common Mistakes and How to Fix Them



Mistake 1: Mapping a variable name outside aes()



❌ geom_point(color = drv) passes the string "drv" as color, coloring all points one color:





  


    
# Wrong: color = drv outside aes() is literal, not a column lookup
ggplot(mpg, aes(x = displ, y = hwy)) + geom_point(color = drv)


    


      

      

    


    


  


  






✅ Move data mappings inside aes():





  


    
ggplot(mpg, aes(x = displ, y = hwy, color = drv)) + geom_point()


    


      

      

    


    


  


  






Mistake 2: Using geom_bar() with pre-computed data



❌ geom_bar() counts rows. Adding y = value causes an error unless you add stat = "identity".



✅ Use geom_col() when you have pre-computed heights.



Mistake 3: Scales applied after geoms that ignore them



❌ scale_x_log10() applied to a histogram changes the axis labels but not the bin boundaries — the histogram was already computed on the original scale.



✅ Log-transform the variable before plotting, or use scale_x_continuous(trans = "log10") which correctly affects stat_bin().



Mistake 4: Forgetting to load ggplot2 or a required extension



❌ patchwork::wrap_plots() fails silently or with a confusing error if patchwork isn't loaded.



✅ library(patchwork) before use. Extensions like ggrepel, patchwork, and ggridges must be explicitly loaded even when ggplot2 is already loaded.



Mistake 5: theme() overriding theme_*() — order matters



❌ Placing theme_minimal() after theme() resets all the custom theme() changes:





  


    
# Wrong order - theme_minimal() resets everything after it
ggplot(mpg, aes(x = displ, y = hwy)) +
  geom_point() +
  theme(axis.text = element_text(color = "red")) +
  theme_minimal()  # this resets the red text!


    


      

      

    


    


  


  






✅ Always put the base theme function before your theme() customizations:





  


    
ggplot(mpg, aes(x = displ, y = hwy)) +
  geom_point() +
  theme_minimal() +  # base theme first
  theme(axis.text = element_text(color = "red"))  # customizations after


    


      

      

    


    


  


  






Practice Exercises



Exercise 1: Build a chart from scratch



Using only this cheat sheet as a reference, create a scatter plot of mpg's cty (x) vs hwy (y) with:



    


  • Points colored by class using scale_color_brewer(palette = "Set2")
    • 


  • A regression line per class (geom_smooth(method="lm", se=FALSE))
    • 


  • Facets by drv using facet_wrap()
    • 


  • theme_minimal() + a bold title
    • 







  


    
# Build it here — use the tables in each section for reference
# ggplot(mpg, aes(x = cty, y = hwy, color = class)) + ...


    


      

      

    


    


  


  






Exercise 2: Recreate a specific chart style



Build a horizontal bar chart of average hwy per manufacturer (top 10 only) that:



    


  • Sorts bars by descending average mpg (fct_reorder + coord_flip())
    • 


  • Uses scale_fill_viridis_d() for fill color
    • 


  • Adds value labels with geom_text(hjust = -0.2)
    • 


  • Removes grid lines with theme(panel.grid.major.y = element_blank())
    • 







  


    
library(forcats)
mfr_avg <- aggregate(hwy ~ manufacturer, mpg, mean)
mfr_top <- head(mfr_avg[order(-mfr_avg$hwy), ], 10)

# Build it here
# ggplot(mfr_top, aes(x = fct_reorder(manufacturer, hwy), y = hwy, fill = manufacturer)) +
#   geom_col(show.legend = FALSE) + coord_flip() + ...


    


      

      

    


    


  


  






Complete Example



This end-to-end example assembles all five sections — geoms, aesthetics, scales, facets, and themes — into a single publication-quality chart:





  


    
# Complete example: ggplot2 grammar in one chart
p_final <- ggplot(
    mpg,
    aes(x = displ, y = hwy, color = hwy, size = cyl)
  ) +
  # Layer 1: scatter points
  geom_point(alpha = 0.7) +
  # Layer 2: trend line per facet
  geom_smooth(method = "lm", se = FALSE,
              color = "grey30", linewidth = 0.7,
              aes(group = 1)) +  # one line per facet, not per color
  # Scales
  scale_color_viridis_c(option = "plasma", name = "Hwy MPG") +
  scale_size_continuous(range = c(1.5, 5), name = "Cylinders") +
  scale_x_continuous(breaks = 1:7, labels = paste0(1:7, "L")) +
  scale_y_continuous(breaks = seq(10, 45, 5)) +
  # Facets
  facet_wrap(
    ~ drv,
    labeller = labeller(drv = c("4"="4-Wheel Drive",
                                 "f"="Front-Wheel Drive",
                                 "r"="Rear-Wheel Drive"))
  ) +
  # Labels
  labs(
    title    = "Engine Displacement vs Fuel Efficiency",
    subtitle = "Points colored by highway MPG; size by cylinder count",
    x        = "Engine Displacement",
    y        = "Highway MPG",
    caption  = "Source: ggplot2::mpg"
  ) +
  # Theme
  theme_minimal(base_size = 12) +
  theme(
    plot.title     = element_text(face = "bold", size = 14),
    plot.subtitle  = element_text(color = "grey50", size = 10),
    strip.text     = element_text(face = "bold"),
    legend.position = "right",
    panel.grid.minor = element_blank()
  )

p_final


    


      

      

    


    


  


  






Summary



ggplot2 grammar in one line: ggplot(data, aes()) + geom_*() + scale_*() + facet_*() + theme_*()



Quick decision guide:







Want to…
Use…






Plot points
geom_point()




Connect points over time
geom_line()




Show frequency counts
geom_bar() or geom_histogram()




Show pre-computed values
geom_col()




Show distribution
geom_boxplot() or geom_violin()




Add trend line
geom_smooth(method="lm")




Add labels
geom_text() or ggrepel::geom_label_repel()




Color by category
aes(color=var) + scale_color_brewer()




Color continuous
aes(color=var) + scale_color_viridis_c()




Custom colors
scale_color_manual(values=c(...))




Split into panels
facet_wrap(~var)




Clean white background
theme_minimal() or theme_classic()




Remove an element
theme(element = element_blank())




Move legend
theme(legend.position = "top")







FAQ



What is the difference between color and fill in ggplot2?



color controls the outline or stroke color — of points, lines, and text. fill controls the interior fill of shapes — bars, polygons, and filled circles (shape 21+). For geom_point() with solid shapes (the default shape 16), only color applies. For geom_col() and geom_histogram(), only fill applies to the bar interior.



How do I save a ggplot2 chart to a file?



Use ggsave(): ggsave("chart.png", plot = p_final, width = 8, height = 5, dpi = 300). ggsave() saves the most recently printed plot if you omit the plot argument. Supported formats: PNG, PDF, SVG, JPEG, TIFF.



How do I combine multiple ggplot2 charts into one figure?



Use the patchwork package: p1 + p2 places charts side-by-side; p1 / p2 stacks them vertically; (p1 | p2) / p3 creates a 2-over-1 layout. Annotate the combined figure with plot_annotation(title = "...").



How do I add a second y-axis in ggplot2?



ggplot2 discourages dual y-axes for good design reasons (they are easily misleading). The supported approach uses scale_y_continuous(sec.axis = sec_axis(~ . / 100, name = "Percent")). The secondary axis must be a mathematical transformation of the primary — you cannot independently scale two different variables.



How do I reorder a categorical axis?



Convert the variable to a factor with levels in the desired order before plotting, or use fct_reorder() from the forcats package inside aes(): aes(x = fct_reorder(category, value)). For frequency ordering, use fct_infreq(category).



References



    


  1. ggplot2 official documentation. https://ggplot2.tidyverse.org/reference/
    2. 


  2. Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis. Springer. https://ggplot2-book.org/
    3. 


  3. Wickham, H. & Grolemund, G. (2017). R for Data Science, Chapter 3. https://r4ds.had.co.nz/data-visualisation.html
    4. 


  4. RStudio/Posit ggplot2 cheat sheet (PDF). https://rstudio.github.io/cheatsheets/html/data-visualization.html
    5. 


  5. Wilke, C. O. (2019). Fundamentals of Data Visualization. O'Reilly. https://clauswilke.com/dataviz/
    6. 





What's Next?



    


  • ggplot2 Getting Started — the full introductory tutorial covering the grammar of graphics from scratch with step-by-step examples.
    • 


  • ggplot2 Scatter Plots — deep dive into geom_point() with color, size, and shape mappings and overplotting solutions.
    • 


  • ggplot2 Themes in R — comprehensive coverage of theme() customization to produce publication-quality charts.
    • 




        


        
      


      

        

        
© 2016- Selva Prabhakaran.
          This work is licensed under the Creative Commons License.