ggplot2 wrapped 2025

Ole! There are so many different ways to look at how your geoms were used across your smashing data visualisations. Maybe the easiest way is a part-to-whole story?

Looking at your top 4 geoms and lumping everything else together…

No surprises but your most used geom was also the geom that appeared in the most distinct files

What a chart full year!
In total 1,096 files were created that contained geoms!

Your busiest month(s) were November.
These months accounted for 19% of your files with geoms!

Things look a lot different thinking about geom diversity!

You got your ggplot2 thesaurus out on:
2025-06-23, 2025-11-25
On these dates you used 9 unique geoms!
geom_area, geom_col, geom_function, geom_label_repel, geom_line, geom_point, geom_rect, geom_segment, geom_textbox, geom_vline

This Venn Diagram shows your top 3 most used geoms and how often you used them in the same code files

That’s kind of interesting!

But wait! You used an solid 32 unique geoms - a Venn Diagram of that size would be a mess!

That’s because we’re not really making a straigh forward plot or chart - we’re visualising a graph! In graphs we are interested in the relationships and connections between things. You might be familiar with hairball-looking graph visualisations - which are just as difficult to extract meaning from as a Venn Diagram.

So, let’s use an UpSet Chart to look at your most inter-connected geoms in 2025

Please move your cursor over the chart below - it’s very interactive.

We’ve thrown away some of your data to make this readable. Geoms only appear if they occur more than 10 times in your data and an interaction has to occur more than once

Oh, wow! Tell me more about UpSet Charts

It’s more than likely you’ve not seen many (or any!) UpSet Charts before. They were only invented in 2012, and the authors quite rightly ask for the original paper to be cited whenever they’re used.

Alexander Lex, Nils Gehlenborg, Hendrik Strobelt, Romain Vuillemot, Hanspeter Pfister. UpSet: Visualization of Intersecting Sets IEEE Transactions on Visualization and Computer Graphics (InfoVis), 20(12): 1983–1992, doi:10.1109/TVCG.2014.2346248, 2014.

On to more fun things!

How can we talk about {ggplot2} without mentioning aesthetics?

All 4,308 times you used a geom they did their magic with either dataviz-level or geom-specific aes

Across all your geoms in 2025, 52% of the time you provided dataviz-level aesthetics.

Ohhh… but that chart mentions different packages?

The 2025 edition of the {ggplot2wrapped} report doesn’t focus on comparing package usage.

But if you take a look at our documentation you can analyse that data yourself

# dataviz-level aesthetic
your_data |> 
  ggplot() +
  aes(x = independent_variable,
      y = dependent_variable) +
  geom_point()

# geom-specific aes
your_data |> 
  ggplot() +
  geom_point(aes(x = independent_variable,
      y = dependent_variable))

Such fussy geoms!

Or such strong dataviz-level aesthetics!

Just like all your data visualisations are different, your geom calls varied a lot!

You might have noticed that’s not all your geoms 👀

That’s because 25% of the time you used geoms with zero arguments!

We’re coming to the end of the report…

Here are some final statistics:

Your longest single geom call was a excellent total of 1,093 characters long. It was geom_richtext from the ggtext.
Your geom with the most geoms was different! It was geom_label_repel with a total of <b>15 arguments</b>.

There’s a lot more you can do with the {ggplot2} package but we’ll sign off with what you were probably expecting and we’d hate to let you down:

This package and report depends on several tools from within R:

Geom usage was analysed using Abstract Syntax Trees via the {astgrepr} package
Static charts were built with {ggplot2}
Interactive charts were built with a mixture of {highcharter} and {upsetjs}
The report itself was written using a Quarto extension called Closeread for scrollytelling

The report was built by me - Charlotte Hadley from gpcds.com.

I can be found on LinkedIn if that’s your thing.

I would genuinely love to see you share your {ggplot2wrapped} reports and charts.

There’s lots more you can do yourself with the add_geom_usage_to_files() function. Share what you build and consider contributing to the package, https://github.com/charliejhadley/ggplot2wrapped.