gglogger gglogger website

gglogger is an R package that logs the calls used to create ggplot2 objects.

This can be useful for debugging, reproducibility, and understanding the sequence of operations used to build a plot.

Have you ever created a plot in ggplot2 and then forgotten how you made it? Or wanted to reproduce a plot but couldn’t remember the exact sequence of operations you used? gglogger can help!

Installation

You can install the gglogger package from source using devtools:

devtools::install_github("pwwang/gglogger")

# or
remotes::install_github("pwwang/gglogger")

Usage

To use the gglogger package, simply load it along with ggplot2 and create your plots as usual. The package will automatically log the calls used to create the plots.

Previous Now
library(ggplot2)

p <- ggplot(mpg, aes(x = displ, y = hwy)) +
    geom_point()
library(ggplot2)
# Just add gglogger after ggplot2
library(gglogger)

p <- ggplot(mpg, aes(x = displ, y = hwy)) +
    geom_point()

# Print the logs
print(p$logs)
## ggplot2::ggplot(mpg, aes(x = displ, y = hwy)) +
##   geom_point()

Evaluate the code in logs to reproduce the plot:

p$logs$evaluate()

You can also attach the variables in an environment for evaluation:

env <- new.env()
env$mpg <- mpg
env$mpg$hwy <- mpg$hwy / 2
p$logs$evaluate(env)

Registering a function from a ggplot2 extension

library(dplyr)
library(gglogger)

mtcars_radar <- mtcars %>%
  as_tibble(rownames = "group") %>%
  mutate_at(vars(-group), scales::rescale) %>%
  tail(4) %>%
  select(1:10)

ggradar <- register(ggradar::ggradar)

p <- ggradar(mtcars_radar, legend.position = "right")
print(p$logs)

# ggradar::ggradar(mtcars_radar, legend.position = "right")

Generating code to reproduce a plot

# p is a ggradar plot created in the previous example

code <- p$logs$gen_code(setup = '
library(dplyr)
library(ggradar)

mtcars_radar <- mtcars %>%
  as_tibble(rownames = "group") %>%
  mutate_at(vars(-group), scales::rescale) %>%
  tail(4) %>%
  select(1:10)
')

cat(code)
## library(dplyr)
## library(ggradar)
##
## mtcars_radar <- mtcars %>%
##   as_tibble(rownames = "group") %>%
##   mutate_at(vars(-group), scales::rescale) %>%
##   tail(4) %>%
##   select(1:10)
##
##
## ggradar::ggradar(mtcars_radar, legend.position = "right")

# eval(parse(text = code)) # to reproduce the plot

Limitations

gglogger cannot log the global settings used to create a plot, such as theme_set(). It can only log the calls used directly to create the plot itself. You may need to manually set these global settings when reproducing a plot, or prepare them using the setup argument in gen_code().

If your data is piped to ggplot function, you need to use |> instead of %>%, otherwise the data will be logged as ..