Automated report

Generating PowerPoint slides using R

About

Sometimes I need to present reports in a PowerPoint format.
This memo shows how to use R to generate simple PPT slides from a dataset.

Introduction

This is a sample analysis of deals per market segment (sectors) over time.
Rendering is done with Quarto and Knitr.
Inspired by a question on the quarto-cli repository.

Download PPT or see github.com/esteveaq/reportppt

Learnings

• functions & parameters
• purrr::map and pmap()
• storing plots in a data frame (tibble)
  

---

Setup

# Load necessary libraries
library(tidyverse)
library(patchwork)
# Set theme
theme_set(theme_minimal())

# Get data 
source("Data/data_deals.R", echo = FALSE)

Parameter and Function

The function contains ggplots.

# Define parameter (categories) to ensure proper slide order on the PPT file
sector.select <- custom_levels

# Define function to create plots with a sector.select variable
make_plots <- function(sector.select) {

# define highlight filter for plot a
data.filt.a <- data.plot.a %>% 
filter(sectors == sector.select)
color.hex <- "#AA2841"

# Plot a
a <-  
  data.plot.a %>% 
  ggplot(mapping = aes(x = fct_reorder(sectors, prop), y = prop)) +
  geom_col(fill = "gray70") +
  #highlight filter
  geom_col(data = data.filt.a, fill = color.hex) + 
  coord_flip() +
  geom_text(aes(label = paste0(prop*100, "%")), 
            vjust = .5,
            hjust = 1.2,
            size = 4,
            color = "white") + #add labels
  scale_y_continuous(labels = scales::percent) + # axis label in %
  labs(x = "Sectors",
       title = paste("Proportion of deals for the",sector.select, "sector"), 
       subtitle = "By number of deals") #relabel 

# define highlight filter for plot b
data.filt.b <- data.plot.b %>% filter(sectors == sector.select)

# Plot b
b <-
  ggplot(data = data.plot.b, aes(x = year, y = n)) +
  geom_col(fill = "gray70") + 
  #highlight filter
  geom_col(data = data.filt.b, fill = color.hex) +
  geom_text(data =  data.filt.b, mapping = aes(label = n), 
            vjust = -0.8, color = color.hex) 

# Assemble plots together using the 'library(patchwork)' syntax
(a + b)
}

Map function and store output

Plots are created for each category and stored in a data frame. Reference : Purrr package

# Define the function that will map 'sector.select' to the function 'make_plots'
plots <- purrr::map(sector.select, make_plots)

# Store generated plots in a tibble 
plots_df <- tibble::tibble(
  sector.select = sector.select,
  plots = plots)

Generate slides

#| output: asis
# Iterate over each row of the 'plots_df' data frame and generate slides using knitr (the markdown function that will generate actual PPT slides)
results_slides <- purrr::pmap_chr(plots_df, function(sector.select, plots) {
  knitr::knit_child(text = c(
    "# Sector `r sector.select`",
    "```{r}",
    "#| echo: false",
    "plots",
    "```",
    ""), envir = environment(), quiet = TRUE)
})

# Print results, and indent with '\n' so Quarto recognizes different slides
cat(results_slides, sep = '\n')

Tip

Can also can be achieved with a parameterized quarto report.