Generate parameterized ppt reports using officer

January 11, 2022

Using code below, I’m able to generate a ppt report for subset of mtcars dataset:

library(ggplot2)
library(tidyverse)
library(patchwork)
library(officer)
library(officedown)
library(glue)

small <- mtcars %>% 
  filter(carb %in% c(1, 2))

p1 <- ggplot(mpg, wt, data = small, colour = cyl)
p2 <- ggplot(mpg, data = small) + ggtitle("small")
p <- p1 | p2

template_pptx <- read_pptx()
report <- add_slide(template_pptx, layout = "Title and Content", master = "Office Theme") %>%
  ph_with(value=p, location = ph_location_label(ph_label = "Content Placeholder 2"))
print(report, target=glue('report for small car.pptx'))

Now let’s say we need to reproduce the report generating process for the following datasets as well:

middle <- mtcars %>% 
  filter(carb %in% c(3, 4))

large <- mtcars %>% 
  filter(carb %in% c(6, 8))

My ideas is to convert multiple ggplots part to a function and save in a script plot.R, then I will write pseudocode script named main.R to run the whole process and generate 3 reports for small, middle, large datasets respectively:

# main.R
for i in c(small, middle, large){
  source('plot.R')
  # maybe need to import and run plot function() from plot.R
  # save figure to ppt
  template_pptx <- read_pptx("./ppt_template.pptx")
  report <- add_slide(template_pptx, layout = "Title and Content", master = "Office Theme") %>%
  ph_with(value=p, location = ph_location_label(ph_label = "Content Placeholder 2"))
  print(report, target=glue('report for {i} car.pptx'))
}

The problem I met is I don’t know how to convert plotting code to a function and pass params (maybe save a config.yaml file in case we have many params?) to the predefined function, and finally generate a parameterized report?

Many thanks for your comments and help at advance.

References:

R: multiple ggplot2 plot using d*ply

https://cran.r-project.org/web/packages/egg/vignettes/Ecosystem.html

R – How to generate parameterized reports via nested tibbles + pwalk(rmarkdown::render)

>Solution :

You could put your plotting code in a function which e.g. takes two arguments, a dataframe (x) and a title.

Similarly put the code to prepare the pptx inside a function, which e.g. takes two arguments, a dataframe (x) and a title or filename or …

In my code below I have put your three datasets in a list and then make use of purrr::iwalk to loop over this list to make a pptx report for each dataset. Using purrr::iwalk the name of the dataset is passed as the second argument to the reporting function.

library(ggplot2)
library(patchwork)
library(dplyr)
library(purrr)
library(officer)
library(glue)

plot_fun <- function(x, title) {
  p1 <- ggplot(data = x, aes(mpg, wt, colour = cyl))
  p2 <- ggplot(data = x, aes(mpg)) + ggtitle(title)
  p1 | p2
}

pptx_fun <- function(x, y) {
  p <- plot_fun(x, title = y)
  template_pptx <- read_pptx()
  report <- add_slide(template_pptx, layout = "Title and Content", master = "Office Theme") %>%
    ph_with(value = p, location = ph_location_label(ph_label = "Content Placeholder 2"))
  print(report, target=glue('report for {y} car.pptx'))
}

data_list <- lapply(list(small = 1:2, medium = 3:4, large = 5:6), function(x) filter(mtcars, carb %in% x))

purrr::iwalk(data_list, pptx_fun)