25 Best ggplot2 Extensions in R: ggrepel, ggforce, ggtext & Beyond

The ggplot2 package powers most R visualizations, but over 100 extension packages expand it with better labels, new chart types, richer themes, and even animation — each one plugging in with the same ggplot() + layer syntax you already know.

Note

Some extensions below are marked "Local only". These packages aren't available in the browser code runner — install them in RStudio to try the code. All other examples run directly in your browser.

How Do You Fix Overlapping Labels and Add Rich Text?

Overlapping labels are the single most common ggplot2 frustration. You add geom_text() to a scatter plot and half the labels pile on top of each other. The ggrepel package fixes this with physics-based repulsion — labels push away from each other and from data points automatically.

1. ggrepel — Non-Overlapping Labels

Let's see ggrepel in action on a scatter plot of car data. Notice how every label is readable without manual positioning.

library(ggplot2) library(ggrepel) # Label the 8 most fuel-efficient cars top_cars <- mtcars[order(-mtcars$mpg)[1:8], ] top_cars$car <- rownames(top_cars) ggplot(mtcars, aes(wt, mpg)) + geom_point(color = "grey60") + geom_point(data = top_cars, color = "steelblue", size = 3) + geom_text_repel(data = top_cars, aes(label = car), size = 3.5, max.overlaps = 20) + labs(title = "Most Fuel-Efficient Cars", x = "Weight (1000 lbs)", y = "Miles per Gallon") + theme_minimal() #> [scatter plot with 8 labeled points, no overlapping text]

Every label connects to its point with a subtle line segment. The max.overlaps argument controls how aggressively labels dodge — higher values allow more labels to appear in crowded regions. Use geom_label_repel() instead for labels with a background box.

Key Insight

ggrepel uses physics-based repulsion to place labels. Each label acts like a charged particle, pushing away from other labels and data points until it finds an open position. This means you never need to manually adjust coordinates.

2. ggtext — Rich Text Formatting in Plots

Standard ggplot2 titles are plain text. With ggtext, you can make individual words bold, colored, or italic using HTML/Markdown syntax right inside your plot titles and annotations.

library(ggtext) ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) + geom_point(size = 2) + labs(title = "Sepal dimensions by setosa, versicolor, and virginica") + scale_color_manual(values = c("#E41A1C", "#377EB8", "#4DAF4A")) + theme_minimal() + theme(plot.title = element_markdown(size = 14), legend.position = "none") #> [scatter plot with colored species names in the title, no legend needed]

The key function is element_markdown() — use it in theme() for any text element (title, subtitle, axis labels, caption). For in-plot annotations, use geom_richtext() which supports the same HTML tags. This technique eliminates the need for a separate legend when you color-code your title.

3. geomtextpath — Curved Text Along Lines (Local only)

geomtextpath places text directly on lines and curves, replacing the need for a separate legend on line charts.

library(geomtextpath) ggplot(economics_long, aes(date, value, color = variable, label = variable)) + geom_textline(size = 4, hjust = 0.8) + theme_minimal() + theme(legend.position = "none") #> [line chart with variable names printed along each curve]

The geom_textline() function replaces geom_line() and bends the text to follow the line's path. Readers see the label right on the data — no legend lookup required.

Try it: Load ggrepel and create a scatter plot of iris using Petal.Length vs Petal.Width. Label only the 5 flowers with the largest Petal.Length using geom_label_repel(). Use ex_top5 for your filtered data.

# Try it: label the top 5 iris flowers by Petal.Length ex_top5 <- iris[order(-iris$Petal.Length)[1:5], ] ggplot(iris, aes(Petal.Length, Petal.Width)) + geom_point(color = "grey70") + # your code here: add highlighted points + geom_label_repel #> Expected: scatter plot with 5 labeled points

Click to reveal solution

ex_top5 <- iris[order(-iris$Petal.Length)[1:5], ] ggplot(iris, aes(Petal.Length, Petal.Width)) + geom_point(color = "grey70") + geom_point(data = ex_top5, color = "tomato", size = 3) + geom_label_repel(data = ex_top5, aes(label = rownames(ex_top5)), size = 3) + theme_minimal() #> [scatter with 5 boxed labels pointing to the largest flowers]

Explanation: geom_label_repel() works like geom_text_repel() but draws a filled rectangle behind each label for better readability.

What Extensions Add Powerful New Geoms?

ggplot2 ships with about 30 geoms (points, lines, bars, etc.), but many common visualizations need specialised geometries. These four packages fill the biggest gaps.

4. ggforce — Zoom, Shapes, and Annotations

ggforce gives you facet_zoom() — a way to zoom into a region of your plot while keeping the full view visible side-by-side.

library(ggforce) ggplot(diamonds[sample(nrow(diamonds), 2000), ], aes(carat, price, color = cut)) + geom_point(alpha = 0.5, size = 1) + facet_zoom(x = carat < 1) + labs(title = "Diamond Prices: Full View + Zoom on Small Stones") + theme_minimal() #> [two-panel plot: left shows all data, right zooms into carat < 1]

The zoomed panel links to the full view with a shaded region, so readers instantly see which portion is enlarged. ggforce also provides geom_mark_ellipse() for annotating clusters, geom_arc_bar() for donut charts, and geom_sina() for jittered violin-style point plots.

Tip

Use facet_zoom() for crowded scatter plots. Instead of filtering data or creating separate plots, facet_zoom gives readers both the overview and the detail in one figure — perfect for presentations where you want to highlight a specific region without losing context.

5. ggridges — Ridgeline Density Plots

When you need to compare distributions across many groups, ridgeline plots stack density curves vertically so every group is visible without faceting.

library(ggridges) ggplot(diamonds[sample(nrow(diamonds), 5000), ], aes(x = price, y = cut, fill = cut)) + geom_density_ridges(alpha = 0.7, scale = 1.2) + labs(title = "Diamond Price Distribution by Cut Quality", x = "Price (USD)", y = "") + theme_minimal() + theme(legend.position = "none") #> [5 stacked density curves, one per cut quality, showing price spread]

The scale parameter controls how much the ridges overlap — values above 1 let them overlap slightly for a compact look. Each ridge is a full density estimate, making it easy to spot bimodal distributions or outlier groups at a glance.

Tip

ggridges shines when you have 5+ groups. With fewer groups, a standard geom_density() with fill works fine. With 10+ groups, ridgelines keep everything readable where facets would need scrolling.

6. ggbeeswarm — Jittered Points That Don't Lie

Standard geom_jitter() places points randomly, which can misrepresent density. ggbeeswarm arranges points in a structured pattern so dense regions look dense and sparse regions look sparse.

library(ggbeeswarm) ggplot(iris, aes(Species, Sepal.Width, color = Species)) + geom_beeswarm(size = 2, cex = 2) + labs(title = "Sepal Width by Species (Beeswarm Layout)") + theme_minimal() + theme(legend.position = "none") #> [three columns of points arranged in bee-swarm pattern]

Unlike jitter, every point position is deterministic — run it twice and you get the same layout. The cex parameter controls point spacing. Use geom_quasirandom() from the same package for a softer, violin-shaped alternative.

7. ggdist — Distribution Visualisation and Raincloud Plots

ggdist combines density plots, dot plots, and interval summaries into single composite visualizations. Its signature output is the "raincloud plot" — a half-violin plus jittered dots plus a summary interval.

library(ggdist) ggplot(iris, aes(x = Species, y = Sepal.Length, fill = Species)) + stat_halfeye(adjust = 0.5, width = 0.6, alpha = 0.7, .width = 0, point_colour = NA) + geom_boxplot(width = 0.15, outlier.shape = NA, alpha = 0.5) + labs(title = "Sepal Length Distribution (Raincloud Style)", y = "Sepal Length (cm)") + theme_minimal() + theme(legend.position = "none") #> [half-violin shapes above narrow boxplots for each species]

The stat_halfeye() function draws a half-density shape, while the layered boxplot adds the familiar median and quartile markers. Set .width = c(0.66, 0.95) to show 66% and 95% credible intervals instead of a boxplot.

Try it: Create a ridgeline plot of airquality$Temp grouped by Month using ggridges. Use factor(Month) for the y-axis and add fill = factor(Month). Store your data prep in ex_aq.

# Try it: ridgeline plot of temperature by month ex_aq <- airquality ex_aq$Month <- factor(ex_aq$Month) ggplot(ex_aq, aes(x = Temp, y = Month)) + # your code here: add geom_density_ridges with fill #> Expected: 5 stacked density curves showing temp distribution per month

Click to reveal solution

ex_aq <- airquality ex_aq$Month <- factor(ex_aq$Month) ggplot(ex_aq, aes(x = Temp, y = Month, fill = Month)) + geom_density_ridges(alpha = 0.6, scale = 1.3) + labs(title = "NYC Temperature by Month", x = "Temperature (F)") + theme_minimal() + theme(legend.position = "none") #> [5 ridges: May (cool, tight) → September (warm, spread)]

Explanation: geom_density_ridges() stacks density estimates vertically. The scale controls overlap, and alpha makes overlapping regions visible.

How Do You Combine Multiple Plots Into One Figure?

Scientific papers and reports almost always need multi-panel figures. Two packages dominate this space, and both work with any ggplot object.

8. patchwork — Combine Plots With Math Operators

patchwork is the most intuitive way to arrange plots. Use + to place plots side by side, / to stack vertically, and | for explicit horizontal layout.

library(patchwork) p1 <- ggplot(mtcars, aes(mpg)) + geom_histogram(bins = 10, fill = "steelblue") + labs(title = "MPG Distribution") + theme_minimal() p2 <- ggplot(mtcars, aes(wt, mpg)) + geom_point() + labs(title = "Weight vs MPG") + theme_minimal() p3 <- ggplot(mtcars, aes(factor(cyl), mpg)) + geom_boxplot(fill = "coral") + labs(title = "MPG by Cylinders") + theme_minimal() (p1 | p2) / p3 + plot_annotation(title = "Motor Trend Car Analysis", tag_levels = "A") #> [2 plots on top row, 1 spanning bottom, labeled A/B/C]

The parentheses control grouping: (p1 | p2) / p3 puts plots 1 and 2 on the top row and plot 3 spanning the full width below. The tag_levels = "A" argument auto-labels each panel as (A), (B), (C) — exactly what journals expect.

Warning

patchwork collects themes from each plot independently. If your panels use different themes, the combined figure can look inconsistent. Apply the same theme_*() to all plots before combining, or use & theme_minimal() after the patchwork expression to apply a theme to every panel at once.

Key Insight

patchwork uses math operators for layout. + and | place plots horizontally, / stacks vertically, and parentheses group sub-layouts. This reads like a formula: (A | B) / C is a 2-over-1 layout.

9. cowplot — Publication-Ready Multi-Panel Layouts

cowplot predates patchwork and offers similar functionality via plot_grid(), plus the ability to add inset plots with draw_plot().

library(cowplot) p_main <- ggplot(mtcars, aes(wt, mpg, color = factor(cyl))) + geom_point(size = 2) + theme_minimal_grid() p_inset <- ggplot(mtcars, aes(factor(cyl))) + geom_bar(fill = "steelblue") + labs(x = "Cylinders", y = "Count") + theme_minimal_hgrid(font_size = 9) ggdraw(p_main) + draw_plot(p_inset, x = 0.6, y = 0.6, width = 0.35, height = 0.35) #> [scatter plot with a small bar chart inset in the top-right corner]

The ggdraw() + draw_plot() pattern places one plot inside another at exact coordinates (0-1 range). This is perfect for overview + detail views. cowplot also provides theme_cowplot() and theme_minimal_grid() — clean themes designed for multi-panel figures.

Try it: Create three plots from iris (histogram of Sepal.Length, scatter of Sepal.Length vs Petal.Length, boxplot of Sepal.Width by Species). Combine them into a 1-row layout with patchwork using | and add panel tags with plot_annotation(tag_levels = "A").

# Try it: 3-panel iris figure with patchwork ex_p1 <- ggplot(iris, aes(Sepal.Length)) + geom_histogram(bins = 15, fill = "steelblue") + theme_minimal() ex_p2 <- ggplot(iris, aes(Sepal.Length, Petal.Length)) + geom_point() + theme_minimal() # your code here: create ex_p3 (boxplot) and combine all three #> Expected: 3 plots side by side, tagged A/B/C

Click to reveal solution

ex_p1 <- ggplot(iris, aes(Sepal.Length)) + geom_histogram(bins = 15, fill = "steelblue") + theme_minimal() ex_p2 <- ggplot(iris, aes(Sepal.Length, Petal.Length)) + geom_point() + theme_minimal() ex_p3 <- ggplot(iris, aes(Species, Sepal.Width, fill = Species)) + geom_boxplot() + theme_minimal() + theme(legend.position = "none") (ex_p1 | ex_p2 | ex_p3) + plot_annotation(tag_levels = "A") #> [three iris plots side by side, tagged (A), (B), (C)]

Explanation: The | operator arranges plots horizontally. plot_annotation(tag_levels = "A") adds sequential letter labels to each panel.

Which Extensions Offer the Best Themes and Color Palettes?

ggplot2 ships with theme_gray(), theme_minimal(), and a handful of others. These five packages give you hundreds of ready-made themes and thousands of color palettes.

10. ggthemes — Ready-Made Publication Themes

ggthemes replicates the visual style of well-known publications and data visualization experts.

library(ggthemes) p_base <- ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) + geom_point(size = 2) p_base + theme_economist() + scale_color_economist() + labs(title = "The Economist Style") #> [scatter plot with blue background, Economist-style axes and fonts]

Other popular options include theme_fivethirtyeight() (bold, minimal), theme_tufte() (maximum data-ink ratio), and theme_wsj() (Wall Street Journal). Each theme comes with a matching scale_color_*() function.

11. hrbrthemes — Typography-First Themes

If you want clean, modern-looking plots with excellent font choices, hrbrthemes delivers with theme_ipsum().

library(hrbrthemes) ggplot(mtcars, aes(wt, mpg)) + geom_point(color = "steelblue", size = 2) + labs(title = "Weight vs Fuel Efficiency", subtitle = "Motor Trend Car Road Tests, 1974", x = "Weight (1000 lbs)", y = "Miles per Gallon") + theme_ipsum() #> [clean plot with Arial Narrow/Roboto Condensed fonts, subtle gridlines]

theme_ipsum() uses a condensed sans-serif font with generous spacing. The result looks professional without any manual theme tweaking. Variants include theme_ipsum_rc() (Roboto Condensed) and theme_ipsum_tw() (Titillium Web).

12. ggsci — Journal and Sci-Fi Color Palettes

ggsci provides color palettes inspired by scientific journals (Lancet, Nature, JAMA), sci-fi franchises, and data tools.

library(ggsci) ggplot(iris, aes(Sepal.Length, Petal.Length, color = Species)) + geom_point(size = 2) + scale_color_npg() + labs(title = "Nature Publishing Group Palette") + theme_minimal() #> [scatter with NPG's distinctive red, blue, green palette]

Other palettes include scale_color_lancet(), scale_color_jama(), scale_color_d3() (D3.js colors), and scale_color_startrek(). Each comes in both scale_color_*() and scale_fill_*() variants.

13. paletteer — One Package, 2,500+ Palettes

Instead of installing 10 palette packages, paletteer gives you unified access to all of them through a single interface.

library(paletteer) ggplot(diamonds[sample(nrow(diamonds), 1000), ], aes(carat, price, color = cut)) + geom_point(alpha = 0.7) + scale_color_paletteer_d("ggsci::nrc_npg") + labs(title = "Accessing NPG Palette via paletteer") + theme_minimal() #> [scatter with NPG colors, accessed through paletteer's unified syntax]

The naming convention is "package::palette_name". Use paletteer_d() for discrete palettes, paletteer_c() for continuous, and paletteer_dynamic() for palettes that adapt to the number of categories. Run palettes_d_names to browse all 2,500+ available palettes.

Tip

paletteer is the meta-package for palettes. One install gives you access to ggsci, viridis, RColorBrewer, wesanderson, and dozens more — all through a consistent syntax.

14. ggnewscale — Two Color Scales in One Plot

By default, ggplot2 only allows one color scale and one fill scale per plot. ggnewscale breaks this limit so you can map different variables to different color scales.

library(ggnewscale) ggplot(mtcars, aes(wt, mpg)) + geom_point(aes(color = hp), size = 3) + scale_color_gradient(low = "lightblue", high = "darkblue", name = "Horsepower") + new_scale_color() + geom_point(data = mtcars[mtcars$mpg > 25, ], aes(color = factor(cyl)), size = 4, shape = 17) + scale_color_manual(values = c("4" = "red", "6" = "orange"), name = "Cylinders\n(efficient cars)") + theme_minimal() #> [scatter with blue gradient for hp, plus red/orange triangles for efficient cars]

The new_scale_color() call resets the color aesthetic so you can assign a fresh scale. Place it between two sets of geom + scale layers. The same pattern works with new_scale_fill().

Try it: Create a scatter plot of mtcars (wt vs mpg) and apply theme_solarized() from ggthemes with scale_color_jama() from ggsci, coloring by factor(cyl). Store your result in ex_themed.

# Try it: combine ggthemes + ggsci ex_themed <- ggplot(mtcars, aes(wt, mpg, color = factor(cyl))) + geom_point(size = 3) # your code here: add theme_solarized() and scale_color_jama() #> Expected: scatter plot with Solarized background and JAMA journal colors

Click to reveal solution

ex_themed <- ggplot(mtcars, aes(wt, mpg, color = factor(cyl))) + geom_point(size = 3) + scale_color_jama() + theme_solarized() + labs(title = "Solarized Theme + JAMA Colors", color = "Cylinders") ex_themed #> [scatter on light yellow Solarized background with JAMA blue/red/green]

Explanation: ggplot2 themes and color scales are independent layers — you can freely mix packages. theme_solarized() controls the background and grid; scale_color_jama() controls the point colors.

How Can You Highlight, Annotate, and Add Statistics?

Sometimes you need to draw the reader's eye to specific data points, add statistical test results, or display correlation patterns. These five extensions handle those needs.

15. gghighlight — Conditional Data Highlighting (Local only)

gghighlight lets you emphasise a subset of data while graying out the rest — no manual filtering needed.

library(gghighlight) ggplot(mtcars, aes(wt, mpg, color = factor(cyl))) + geom_point(size = 3) + gghighlight(mpg > 25) + labs(title = "Highlighting Fuel-Efficient Cars (MPG > 25)") + theme_minimal() #> [all points visible, but only mpg > 25 points are colored; rest are grey]

The gghighlight() function works with any geom — scatter, line, bar. It accepts any logical condition and automatically fades non-matching data to grey. Use label_key = car_name to auto-label the highlighted points.

16. ggpubr — Publication-Ready Plots With Statistical Tests

ggpubr wraps common plot types in convenience functions and adds built-in statistical comparisons.

library(ggpubr) ggboxplot(iris, x = "Species", y = "Sepal.Length", color = "Species", palette = "jco", add = "jitter", shape = "Species") + stat_compare_means(comparisons = list( c("setosa", "versicolor"), c("versicolor", "virginica"), c("setosa", "virginica")), method = "t.test") #> [boxplots with jittered points and p-value brackets between groups]

The stat_compare_means() function runs statistical tests (t-test, Wilcoxon, ANOVA, Kruskal-Wallis) and displays the p-value directly on the plot. The comparisons argument specifies which pairs to test. The palette = "jco" uses Journal of Clinical Oncology colors.

17. ggcorrplot — Correlation Matrix Heatmaps

ggcorrplot turns a correlation matrix into a publication-ready heatmap with one function call.

library(ggcorrplot) cor_mat <- round(cor(mtcars[, c("mpg", "cyl", "disp", "hp", "drat", "wt", "qsec")]), 2) ggcorrplot(cor_mat, method = "circle", type = "lower", lab = TRUE, lab_size = 3, colors = c("#6D9EC1", "white", "#E46726")) + labs(title = "Motor Trend Cars: Variable Correlations") #> [lower-triangle correlation matrix with circles sized by correlation strength]

The method = "circle" uses circle size to encode correlation strength (large circle = strong correlation). Set type = "lower" to show only the lower triangle, avoiding redundancy. The lab = TRUE overlays the actual correlation values.

Key Insight

Correlation heatmaps reveal multicollinearity at a glance. When building regression models, look for variable pairs with correlations above 0.8 — including both in a model inflates standard errors and makes coefficients unreliable.

18. GGally — Scatter Plot Matrices and More

GGally's ggpairs() creates comprehensive scatter-plot matrices showing correlations, distributions, and scatter plots for every variable pair.

library(GGally) ggpairs(iris, columns = 1:4, aes(color = Species, alpha = 0.5), upper = list(continuous = wrap("cor", size = 3)), lower = list(continuous = wrap("points", size = 0.5)), diag = list(continuous = wrap("densityDiag", alpha = 0.5))) + theme_minimal() + labs(title = "Iris: All Pairwise Relationships") #> [4x4 matrix: correlations above diagonal, scatters below, densities on diagonal]

The upper triangle shows correlation coefficients, the lower triangle shows scatter plots, and the diagonal shows density distributions — all colored by Species. This single chart replaces six separate scatter plots and gives you a complete overview of multivariate relationships.

19. ggstatsplot — Automated Statistical Visualisations (Local only)

ggstatsplot creates publication-ready plots with statistical test results embedded directly in the subtitle.

library(ggstatsplot) ggbetweenstats(iris, x = Species, y = Sepal.Length, type = "parametric", pairwise.display = "significant") #> [violin + box + jitter plot with ANOVA F-statistic, p-value, #> effect size, and pairwise comparisons all printed on the plot]

One function call runs the appropriate test, computes effect sizes, performs pairwise comparisons, and displays everything on the plot. Other variants include ggscatterstats() (correlation), gghistostats() (distribution tests), and ggcorrmat() (correlation matrices with p-values).

Warning

ggstatsplot auto-selects tests based on your data. It chooses between parametric and non-parametric tests, but always verify the chosen test matches your assumptions. Set type = "parametric" or type = "nonparametric" explicitly when you know which is appropriate.

Try it: Use ggcorrplot() to create a correlation heatmap of airquality (columns: Ozone, Solar.R, Wind, Temp). Use method = "circle", type = "lower", and add labels. Store the correlation matrix in ex_cor.

# Try it: airquality correlation heatmap ex_aq_clean <- na.omit(airquality[, c("Ozone", "Solar.R", "Wind", "Temp")]) ex_cor <- round(cor(ex_aq_clean), 2) # your code here: call ggcorrplot with circle method, lower triangle, labels #> Expected: lower-triangle circle heatmap with correlation values

Click to reveal solution

ex_aq_clean <- na.omit(airquality[, c("Ozone", "Solar.R", "Wind", "Temp")]) ex_cor <- round(cor(ex_aq_clean), 2) ggcorrplot(ex_cor, method = "circle", type = "lower", lab = TRUE, lab_size = 3.5, colors = c("#6D9EC1", "white", "#E46726")) + labs(title = "Air Quality: Variable Correlations") #> [lower-triangle heatmap: Ozone-Temp strongly positive, Wind-Temp negative]

Explanation: na.omit() removes rows with missing values before computing correlations. The circle sizes immediately show that Ozone and Temp have the strongest positive correlation.

How Do You Make ggplot2 Charts Interactive, Animated, or Specialised?

The final group covers extensions that take ggplot2 beyond static PNG output — plus three packages for niche chart types you can't build with base geoms.

20. gganimate — Animated Plots (Local only)

gganimate adds a time dimension to any ggplot by treating "frame" as an aesthetic. It requires the gifski package for rendering GIFs — install both with install.packages(c("gganimate", "gifski")).

library(gganimate) ggplot(gapminder::gapminder, aes(gdpPercap, lifeExp, size = pop, color = continent)) + geom_point(alpha = 0.7) + scale_x_log10() + transition_time(year) + labs(title = "Year: {frame_time}", x = "GDP per Capita (log)", y = "Life Expectancy") + theme_minimal() #> [animated scatter plot cycling through years 1952-2007]

Key transition functions: transition_time() for temporal data, transition_states() for categorical groups, and transition_reveal() for cumulative line charts. Use shadow_wake() to leave trails behind moving points.

21. plotly — Interactive HTML Charts With ggplotly() (Local only)

plotly's ggplotly() function converts any ggplot into an interactive chart with zoom, pan, and hover tooltips. Install locally with install.packages("plotly").

library(plotly) p <- ggplot(iris, aes(Sepal.Length, Petal.Length, color = Species, text = paste("Width:", Sepal.Width))) + geom_point(size = 2) + theme_minimal() ggplotly(p, tooltip = c("text", "x", "y")) #> [interactive scatter: hover shows species, dimensions; scroll to zoom]

The tooltip argument controls what appears on hover. Map extra information to the text aesthetic and include it in tooltip for rich hover cards. For dashboards, combine with Shiny or save standalone HTML with htmlwidgets::saveWidget().

22. treemapify — Treemap Charts

treemapify turns hierarchical or proportional data into space-filling rectangles.

library(treemapify) library(dplyr) tree_data <- mtcars |> mutate(car = rownames(mtcars), cyl = factor(cyl)) |> group_by(cyl) |> summarise(count = n(), avg_mpg = mean(mpg), .groups = "drop") ggplot(tree_data, aes(area = count, fill = avg_mpg, label = paste0(cyl, " cyl\n", count, " cars"))) + geom_treemap() + geom_treemap_text(place = "centre", size = 12) + scale_fill_gradient(low = "#E46726", high = "#6D9EC1", name = "Avg MPG") + labs(title = "Cars by Cylinder Count (sized by count, colored by efficiency)") + theme_minimal() #> [three rectangles: 8-cyl largest (14 cars), 4-cyl (11), 6-cyl (7)]

Map area to the variable that controls rectangle size, and fill to a second variable for color encoding. Use geom_treemap_text() to label each rectangle. For hierarchical treemaps, add subgroup aesthetics.

23. waffle — Waffle Charts (Square Pie Charts)

Waffle charts display proportions as grids of squares — a more readable alternative to pie charts.

library(waffle) parts <- c("4 cyl" = 11, "6 cyl" = 7, "8 cyl" = 14) ggplot(data.frame(category = names(parts), count = parts), aes(fill = category, values = count)) + geom_waffle(n_rows = 4, size = 0.5, color = "white") + scale_fill_manual(values = c("#4DAF4A", "#377EB8", "#E41A1C")) + coord_equal() + labs(title = "Cars by Cylinder Count") + theme_void() + theme(legend.position = "bottom") #> [grid of colored squares: 11 green, 7 blue, 14 red, in rows of 4]

Each square represents one unit. The n_rows parameter sets how many rows tall the grid is, and coord_equal() keeps the squares square. Waffle charts work best when your total count is under 100 — otherwise each square is too small to see.

Tip

Waffle charts beat pie charts for comparing proportions. Human brains estimate area in rectangles more accurately than in circle slices. Use waffle charts whenever you'd reach for a pie chart — your readers will thank you.

24. ggiraph — Interactive SVG Graphics (Local only)

ggiraph makes ggplot elements interactive — hover for tooltips, click to select, and link data across plots.

library(ggiraph) p <- ggplot(mtcars, aes(wt, mpg)) + geom_point_interactive(aes(tooltip = rownames(mtcars), data_id = rownames(mtcars)), size = 3, color = "steelblue") + theme_minimal() girafe(ggobj = p) #> [scatter plot where hovering shows car name, clicking highlights the point]

Replace standard geoms with _interactive versions (geom_point_interactive, geom_bar_interactive, etc.) and map tooltip and data_id aesthetics. The data_id enables linked brushing — clicking a point in one chart highlights the same observation in another.

25. ggalluvial — Alluvial and Sankey Diagrams (Local only)

ggalluvial visualises flow between categorical variables — tracking how observations move across stages.

library(ggalluvial) titanic_data <- as.data.frame(Titanic) ggplot(titanic_data, aes(axis1 = Class, axis2 = Sex, axis3 = Survived, y = Freq)) + geom_alluvium(aes(fill = Survived), width = 1/3) + geom_stratum(width = 1/3) + geom_text(stat = "stratum", aes(label = after_stat(stratum)), size = 3) + scale_fill_manual(values = c("No" = "#E41A1C", "Yes" = "#4DAF4A")) + labs(title = "Titanic Survival by Class and Sex") + theme_void() #> [flowing ribbons showing how passengers in each class/sex survived or not]

The geom_alluvium() draws flowing ribbons between strata, and geom_stratum() draws the category blocks. This chart type works best for 3-5 categorical variables with limited categories each (under 6 per axis).

Try it: Create a treemap of the diamonds dataset showing proportions by cut. Group by cut, count the rows in each group, and map area = n and fill = cut. Use geom_treemap() and geom_treemap_text().

# Try it: diamonds treemap by cut ex_cuts <- diamonds |> group_by(cut) |> summarise(n = n(), .groups = "drop") # your code here: ggplot + geom_treemap + geom_treemap_text #> Expected: 5 rectangles sized by diamond count, labeled with cut names

Click to reveal solution

ex_cuts <- diamonds |> group_by(cut) |> summarise(n = n(), .groups = "drop") ggplot(ex_cuts, aes(area = n, fill = cut, label = paste0(cut, "\n", n))) + geom_treemap() + geom_treemap_text(place = "centre", size = 11) + labs(title = "Diamonds by Cut Quality") + theme_minimal() + theme(legend.position = "none") #> [Ideal largest rectangle (21,551), then Premium, Very Good, Good, Fair]

Explanation: geom_treemap() sizes rectangles proportionally by the area aesthetic. Ideal cut dominates because it has the most diamonds in the dataset.

Practice Exercises

Exercise 1: Multi-Panel Figure With Labels and Ridgelines

Create a two-panel figure using patchwork: (A) a scatter plot of mtcars (wt vs mpg) with the 5 most fuel-efficient cars labeled using ggrepel, and (B) a ridgeline plot of mpg by factor(cyl) using ggridges. Apply theme_minimal() to both and add panel tags.

# Exercise 1: combine ggrepel + ggridges + patchwork # Hint: create two separate ggplot objects, then use | and plot_annotation()

Click to reveal solution

my_top5 <- mtcars[order(-mtcars$mpg)[1:5], ] my_top5$car <- rownames(my_top5) my_p1 <- ggplot(mtcars, aes(wt, mpg)) + geom_point(color = "grey60") + geom_point(data = my_top5, color = "steelblue", size = 3) + geom_text_repel(data = my_top5, aes(label = car), size = 3) + labs(title = "Weight vs MPG", x = "Weight", y = "MPG") + theme_minimal() my_p2 <- ggplot(mtcars, aes(x = mpg, y = factor(cyl), fill = factor(cyl))) + geom_density_ridges(alpha = 0.6) + labs(title = "MPG by Cylinders", x = "MPG", y = "Cylinders") + theme_minimal() + theme(legend.position = "none") (my_p1 | my_p2) + plot_annotation(tag_levels = "A") #> [two-panel figure: (A) labeled scatter, (B) ridgeline densities]

Explanation: Each plot is built independently with its own extension, then combined with | for side-by-side layout. plot_annotation(tag_levels = "A") labels them (A) and (B).

Exercise 2: Styled Correlation Matrix

Build a correlation matrix of mtcars columns (mpg, hp, wt, qsec, drat) using ggcorrplot. Use method = "square" and type = "upper". Apply a Nature Publishing Group (scale_fill_npg() won't work here — use custom colors from ggsci's NPG palette). Add a ggtext-styled title where "positive" is blue and "negative" is red.

# Exercise 2: ggcorrplot + ggsci-inspired colors + ggtext title # Hint: use ggcorrplot's colors argument for custom palette # Hint: use labs(title = ...) with HTML spans, then element_markdown()

Click to reveal solution

my_cor <- round(cor(mtcars[, c("mpg", "hp", "wt", "qsec", "drat")]), 2) ggcorrplot(my_cor, method = "square", type = "upper", lab = TRUE, lab_size = 4, colors = c("#3C5488", "white", "#E64B35")) + labs(title = "Positive and negative correlations in mtcars") + theme(plot.title = element_markdown(size = 13)) #> [upper-triangle square heatmap with NPG-inspired blue/red colors #> and a title with colored words]

Explanation: ggcorrplot's colors argument takes three values (low, mid, high). The NPG blue (#3C5488) and red (#E64B35) are from ggsci's palette. element_markdown() from ggtext renders the HTML spans in the title.

Exercise 3: Zoom + Rich Labels + Custom Palette

Create a scatter plot of diamonds (carat vs price, colored by clarity, sampled to 2000 rows) with: (1) facet_zoom(x = carat < 0.5) from ggforce to zoom into small diamonds, (2) a ggtext-styled subtitle that highlights the word "small" in bold red, and (3) a ggsci scale_color_d3() palette. Save as my_zoom_plot.

# Exercise 3: ggforce + ggtext + ggsci # Hint: use element_markdown() for the subtitle # Hint: sample diamonds first for speed

Click to reveal solution

set.seed(99) my_diamonds <- diamonds[sample(nrow(diamonds), 2000), ] my_zoom_plot <- ggplot(my_diamonds, aes(carat, price, color = clarity)) + geom_point(alpha = 0.5, size = 1.5) + facet_zoom(x = carat < 0.5) + scale_color_d3() + labs(title = "Diamond Prices by Clarity", subtitle = "Zooming into small stones (< 0.5 carat)") + theme_minimal() + theme(plot.subtitle = element_markdown()) my_zoom_plot #> [two-panel zoom: left = all data with D3 colors, right = carat < 0.5 zoomed]

Explanation: facet_zoom() creates a linked zoom panel. The ggtext subtitle uses HTML to make "small" bold and red. scale_color_d3() provides a distinctive 10-color palette from the D3.js library.

Putting It All Together

Let's build a polished multi-panel analysis of mtcars that combines five extensions in one figure: ggrepel for labels, ggtext for a rich title, ggsci for a professional palette, ggcorrplot for correlations, and patchwork to combine everything.

# Panel A: Scatter with labeled outliers top_efficient <- mtcars[order(-mtcars$mpg)[1:5], ] top_efficient$car <- rownames(top_efficient) panel_a <- ggplot(mtcars, aes(wt, mpg, color = factor(cyl))) + geom_point(size = 2.5) + geom_text_repel(data = top_efficient, aes(label = car), color = "black", size = 3, max.overlaps = 10) + scale_color_npg(name = "Cylinders") + labs(x = "Weight (1000 lbs)", y = "MPG") + theme_minimal() # Panel B: Correlation matrix cor_vars <- round(cor(mtcars[, c("mpg", "hp", "wt", "disp", "drat")]), 2) panel_b <- ggcorrplot(cor_vars, method = "circle", type = "lower", lab = TRUE, lab_size = 3, colors = c("#3C5488", "white", "#E64B35")) # Combine with patchwork (panel_a | panel_b) + plot_annotation( title = "Motor Trend Cars: Performance vs Efficiency", tag_levels = "A", theme = theme(plot.title = element_markdown(size = 15)) ) #> [A: scatter with NPG colors and ggrepel labels | B: correlation circles] #> [Overall title with blue "Performance" and red "Efficiency"]

This figure tells a complete story: Panel A reveals the weight-efficiency trade-off with standout cars labeled, while Panel B quantifies how each variable relates to the others. The NPG colors and ggtext title make it publication-ready. Five extensions, one unified figure — that's the power of the ggplot2 ecosystem.

Summary

Here's every extension covered in this article, with its category, key function, and whether it runs in the browser code blocks.

#	Package	Category	Key Function	Browser?
1	ggrepel	Labels	`geom_text_repel()`	Yes
2	ggtext	Labels	`element_markdown()`	Yes
3	geomtextpath	Labels	`geom_textline()`	No
4	ggforce	Geoms	`facet_zoom()`	Yes
5	ggridges	Geoms	`geom_density_ridges()`	Yes
6	ggbeeswarm	Geoms	`geom_beeswarm()`	Yes
7	ggdist	Geoms	`stat_halfeye()`	Yes
8	patchwork	Composition	`+`, `/`, `	` operators	Yes
9	cowplot	Composition	`plot_grid()`, `draw_plot()`	Yes
10	ggthemes	Themes	`theme_economist()`	Yes
11	hrbrthemes	Themes	`theme_ipsum()`	Yes
12	ggsci	Palettes	`scale_color_npg()`	Yes
13	paletteer	Palettes	`scale_color_paletteer_d()`	Yes
14	ggnewscale	Palettes	`new_scale_color()`	Yes
15	gghighlight	Annotation	`gghighlight()`	No
16	ggpubr	Statistics	`stat_compare_means()`	Yes
17	ggcorrplot	Statistics	`ggcorrplot()`	Yes
18	GGally	Statistics	`ggpairs()`	Yes
19	ggstatsplot	Statistics	`ggbetweenstats()`	No
20	gganimate	Animation	`transition_time()`	No
21	plotly	Interactive	`ggplotly()`	No
22	treemapify	Specialised	`geom_treemap()`	Yes
23	waffle	Specialised	`geom_waffle()`	Yes
24	ggiraph	Interactive	`geom_point_interactive()`	No
25	ggalluvial	Specialised	`geom_alluvium()`	No

Every extension follows the same pattern: install the package, load it with library(), and add its geoms, themes, or scales to your existing ggplot code. Start with the ones that solve your most pressing pain point — for most people, that's ggrepel (labels) and patchwork (multi-panel figures).

References

Official ggplot2 Extensions Gallery — curated list of 100+ community packages. Link
Wickham, H. — ggplot2: Elegant Graphics for Data Analysis, 3rd Edition. Springer (2024). Link
Slowkowski, K. — ggrepel documentation: repulsive text and label geoms. Link
Pedersen, T.L. — patchwork: the composer of plots. Link
Pedersen, T.L. — ggforce: accelerating ggplot2. Link
Wilke, C.O. — ggtext: improved text rendering for ggplot2. Link
Hvitfeldt, E. — paletteer: comprehensive collection of colour palettes. Link
Pedersen, T.L. & Robinson, D. — gganimate: a grammar of animated graphics. Link
Kassambara, A. — ggpubr: publication-ready plots. Link
Patil, I. — ggstatsplot: ggplot2-based plots with statistical details. Link

Continue Learning

ggforce Package in R — Deep dive into facet_zoom, mark ellipses, arc bars, sina plots, and 30+ advanced ggplot2 geoms
patchwork in R — Master multi-panel figures with aligned axes, shared legends, and complex nested layouts
ggdist Package in R — Raincloud plots, half-eye plots, uncertainty bands, and distribution visualisation with stat_halfeye and stat_dots

r‑statistics.co by Selva Prabhakaran

25 Best ggplot2 Extensions in R: ggrepel, ggforce, ggtext & Beyond

How Do You Fix Overlapping Labels and Add Rich Text?

1. ggrepel — Non-Overlapping Labels

2. ggtext — Rich Text Formatting in Plots

3. geomtextpath — Curved Text Along Lines (Local only)

What Extensions Add Powerful New Geoms?

4. ggforce — Zoom, Shapes, and Annotations

5. ggridges — Ridgeline Density Plots

6. ggbeeswarm — Jittered Points That Don't Lie

7. ggdist — Distribution Visualisation and Raincloud Plots

How Do You Combine Multiple Plots Into One Figure?

8. patchwork — Combine Plots With Math Operators

9. cowplot — Publication-Ready Multi-Panel Layouts

Which Extensions Offer the Best Themes and Color Palettes?

10. ggthemes — Ready-Made Publication Themes

11. hrbrthemes — Typography-First Themes

12. ggsci — Journal and Sci-Fi Color Palettes

13. paletteer — One Package, 2,500+ Palettes

14. ggnewscale — Two Color Scales in One Plot

How Can You Highlight, Annotate, and Add Statistics?

15. gghighlight — Conditional Data Highlighting (Local only)

16. ggpubr — Publication-Ready Plots With Statistical Tests

17. ggcorrplot — Correlation Matrix Heatmaps

18. GGally — Scatter Plot Matrices and More

19. ggstatsplot — Automated Statistical Visualisations (Local only)

How Do You Make ggplot2 Charts Interactive, Animated, or Specialised?

20. gganimate — Animated Plots (Local only)

21. plotly — Interactive HTML Charts With ggplotly() (Local only)

22. treemapify — Treemap Charts

23. waffle — Waffle Charts (Square Pie Charts)

24. ggiraph — Interactive SVG Graphics (Local only)

25. ggalluvial — Alluvial and Sankey Diagrams (Local only)

Practice Exercises

Exercise 1: Multi-Panel Figure With Labels and Ridgelines

Exercise 2: Styled Correlation Matrix

Exercise 3: Zoom + Rich Labels + Custom Palette

Putting It All Together

Summary

References

Continue Learning

Related Tutorials

On this page