Skip to contents

ggplot2wrapped

Your year in {ggplot2}!

Check out an example ggplot2wrapped report for 2025!

This package attempts to give you the Spotify Wrapped experience but looking at how you use the incredible {gglot2} package for visualising your data!

Within the package you will find 3 kinds of things:

  • ggplot_wrapped_2025() for building your personalised interactive report on your year in {ggplot2}

  • Functions for extracting your geom usage from across your whole code base.

  • Individual visualisation functions used within the interactive report for exploring your geom usage

This runs fully local and doesn’t use LLMs. In fact, not only is it LLM free it was also built and designed by hand. Slop free through and through.

It provides lots of visualisations and analysis tools for your ggplot2 usage!

Installation

This package is currently only available on GitHub and has a few dependencies to be aware of.

  • You need to install remotes::install_github("charliejhadley/GPCDStools")

This should then work:

remotes::install_github("charliejhadley/ggplot2wrapped")

This package depends on {astgrepr} at the moment. It’s available on CRAN, but if you are on Linux or have other exotic setups you might need to faff a little to get the dependency setup.

Take a minute to read about bugs in this package.

Generating your own report

First you need to collect info about your code files. Modify the first argument to look at the locations you’re interested in looking at.

There is example usage data in the data_example_geom_usage object if you prefer

library("ggplot2wrapped")
code_files <- get_code_file_info(c("~/Github/", "~/coding/r-projects-scrapbook/"), file_type = c(".qmd", ".R", ".Rmd"))

code_files
#> # A tibble: 2 × 4
#>   file_path          modified_time       created_time        access_time        
#>   <chr>              <dttm>              <dttm>              <dttm>             
#> 1 /Users/charliejha… 2022-04-20 20:10:12 2022-04-20 20:10:12 2025-11-21 14:28:18
#> 2 /Users/charliejha… 2022-04-20 20:10:14 2022-04-20 20:10:14 2025-11-21 14:06:55

Great! Now this package was thrown together in a week so makes a few assumptions and the report assumes you’re looking at 2025 data - so let’s filter that data.

library("tidyverse")
target_year <- 2025
code_files_from_target_year <- code_files |>
  filter(year(modified_time) == target_year)

Now the clever function add_geom_usage_to_files() extracts your geom usage, if you have lots of files it might take a few minutes. Which geoms do we look at?

The output of this function contains potentially sensitive information!!

Your actual geom calls are in the function_call argument so it might say geom_point(size = state_secrets).

When you build the {ggplot2wrapped} nice report it will drop this argument - no sensitive information is included in the report.

data_geoms_usage_target_year <- code_files_from_target_year |> 
  add_geom_usage_to_files()
#> # A tibble: 2 × 12
#>   file_path               file_extension modified_time       created_time       
#>   <chr>                   <chr>          <dttm>              <dttm>             
#> 1 /Users/charliejhadley/… R              2021-08-31 17:53:54 2022-04-07 19:50:42
#> 2 /Users/charliejhadley/… R              2021-08-31 17:53:54 2022-04-07 19:50:42
#> # ℹ 8 more variables: access_time <dttm>, geom_name <chr>, has_aes <lgl>,
#> #   function_call <list>, length_of_call <dbl>, n_args_in_call <dbl>,
#> #   n_times_used <int>, package_name <chr>

Neat! You’re now ready to create your own report, it’ll be saved into your project/working directory with the name Charlie's ggplot2wrapped.html

There is example usage data in the data_example_geom_usage object if you prefer

data_geoms_usage_target_year %>% 
  ggplot_wrapped_2025(data_geoms, report_year = target_year, exported_report_name = "Charlie's ggplot2wrapped.html")

Enjoy reading and please consider sharing your report with me (Charlotte Hadley) as I’d love to see it!

Which geoms do we look at?

Ah. This package uses the {astgrepr} to work with the Abstract Syntax Trees of your code files. We need to be able to differentiate between these:

# We want
ggplot() +
  geom_point()

# We don't want
g_title <- "I love geom_point"

# We don't want
geom_state_secret_custom <- function(){
  
}

We therefore explicitly include geoms with an allowlist to be included in the analysis. Earlier there was a default argument used that we didn’t mention:

data_geoms_usage_target_year <- code_files_from_target_year |> 
  add_geom_usage_to_files(data_geoms = ggplot2wrapped::data_geoms)

The data_geoms dataset includes all geoms from {ggplot2}, {ggrepel} and {ggtext}. You’re very welcome to provide a large white list, but note again that custom geom names and their arguments might be sensitive.

What visualisations do I get?

We gave ourselves a week to write this package, the report and the visualisations within. There are currently 10 visualisations available.

Almost of these are available as functions within the package and can be used as below. Note that we’ll be increasing the customisation of these functions after we release the 2025 wrapped report!

data_example_geom_usage |> 
  make_geom_interactive_upset_chart(height = 600)

Side projects and bugs

Side projects almost never lead anywhere. I’ve committed to actually making something from ideas so when I made the first commit on Friday 21st Dec I promised myself I’d launch a workable package on December 1st.

And it’s here! But also, there will be bugs. That’s what happens when your sprints refuse to look backwards.

Please report any issues. If the fancy ggplot_wrapped_2025() function doesn’t work then try out the visualisations or analyse your data returned by add_geom_usage_to_files()