parameterizedPPT.qmd

---
title: "Automating PPT reports with Quarto"
subtitle: "Providing feedback from survey data to multiple groups"
author:
  name: 'Magnus Johansson'
  affiliation: 'RISE Research Institutes of Sweden'
  affiliation-url: 'https://ri.se/shic'
  orcid: '0000-0003-1669-592X'
date: last-modified
date-format: iso
google-scholar: true
citation:
  type: 'webpage'
format:
  html:
    css: styles.css
execute: 
  cache: false
  warning: false
  message: false
editor_options: 
  chunk_output_type: console
---

## Background

In a recent study we wanted to provide group-level feedback to all participants from the baseline survey.

While I can't show all the parts of this, I will simulate some data and hopefully we will end up with something that is useful and can be generalized from.

## Setting up

Things that are needed:

- dataframe with participant data
- df with information about our items
- a flexible function for plotting (theming optional)
- Quarto (.qmd) file with the grouping parameter
  - this file also imports/recodes data, reads item info and the plotting function
  - here you create the structure needed for the PPT
- R-file with code for rendering multiple PPT-files based on the template .qmd

### Item info

We'll need a dataframe with information about each item:

- short name/code
- full item text
- response categories
- intro text
- domain/questionnaire, if some items belong together

For this, we'll use WHO-5 as an example.

```{r}

```

Our dataframe looks like this:

```{r}

```

In our case, there is a separate file where participant ID's can be matched to a group, which can be joined to the dataset (if you don't collect group info in the survey).

## Defining a plot function

First we need to decide which things we will want to be able to adjust - options in the function.

It is good practice to offer the dataframe as first input. Then we have the items, followed by intro text, response categories, and number of faceting rows.

```{r}

# ändra så att svarskategorier hämtas från itemlabels-objektet
staplar <- function(data, items, rubrik, rader) {
  
  # get response categories to a vector
  svarskategorier <- data %>% 
    select(all_of({{ items }})) %>% 
    pivot_longer(everything()) %>%
    na.omit() %>% 
    distinct(value) %>% 
    pull(value)
  
  p <- data %>% 
    select(all_of({{ items }})) %>% 
    pivot_longer(everything()) %>%
    na.omit() %>% 
    count(name, value) %>% # räkna hur många individer i varje svarskategori
    mutate(Procent = (100 * n / sum(n))) %>% # räkna fram procent för varje svarskategori
    mutate(across(where(is.numeric), ~ round(.x, 2))) %>%
    rename(
      Svarskategori = value, # byt namn på variabler inför skapande av figur.
      Fråga = name
    ) %>%
    left_join(itemlabels, by = join_by("Fråga" == "itemnr")) %>% 
    ggplot(aes(x = Svarskategori, y = Procent, fill = Svarskategori)) +
    scale_y_continuous(limits = c(0,NA)) + # starta alltid y på 0
    scale_x_discrete(labels = c(0:length(svarskategorier))) + # byt etikett på x axeln så det blir mindre text i figuren
    geom_col() +
    geom_text(aes(label = n), color = "darkgrey",
              size = 4,
              position = position_dodge(width = 0.9),
              vjust = -0.24) +
    theme_rise_ppt() +
    scale_fill_viridis_d(labels = ~ str_wrap(.x, width = 9)) + #radbrytning i förklaringsrutan
    theme(panel.grid.major.x = element_blank(),
          panel.grid.minor.x = element_blank()) +
    labs(title = rubrik, 
         caption = str_wrap("Staplarnas höjd anger andel (%) av respondenterna som svarat i respektive svarskategori. Siffran ovanför staplarna anger antalet respondenter som svarat i den kategorin.", 100))
  
  if (missing(rader) == TRUE) {
    p + facet_wrap(~item, 
                   labeller = label_wrap_gen(22)) 
  }
  else {
    p + facet_wrap(~item, 
                   nrow = rader,
                   labeller = label_wrap_gen(22)) 
  }
}
```

### Theming

```{r}
theme_rise_ppt <- function(fontfamily = "Lato", axissize = 17, titlesize = 18,
                       margins = 12, axisface = "plain", stripsize = 15, axisTsize = 15, capsize = 14,
                       panelDist = 0.6, legendSize = 15, legendTsize = 16, ...) {
  theme_minimal() +
    theme(
      text = element_text(family = fontfamily),
      axis.title.x = element_text(
        margin = margin(t = margins),
        size = axissize
      ),
      axis.title.y = element_text(
        margin = margin(r = margins),
        size = axissize
      ),
      plot.title = element_text(
        face = "bold",
        size = titlesize
      ),
      axis.title = element_text(
        face = axisface
      ),
      plot.caption = element_text(
        face = "italic", size = capsize
      ),
      axis.text = element_text(size = axisTsize),
      legend.text = element_text(family = fontfamily, size = legendSize),
      legend.title = element_text(family = fontfamily, size = legendTsize),
      legend.background = element_rect(color = "lightgrey"),
      strip.text = element_text(size = stripsize),
      strip.background = element_rect(color = "lightgrey"),
      panel.spacing = unit(panelDist, "cm", data = NULL),
      ...
    )
}
```


## Quarto parameters

We only have one parameter in this example - "group".

## Render code

This is a separate .R file 

```{r}
library(glue)
library(quarto)
library(purrr)
library(readxl)
library(dplyr)
library(janitor)

### Credit to https://github.com/Pecners/sra_pullout/blob/main/render.R for inspiration

# calculate no of responses in each group and identify those with low number of responses (<6)


# get group names
groups <- read_excel("data/AFA_mailadresser.xlsx", sheet = 1) %>% 
  clean_names() %>% 
  distinct(group) %>% 
  filter(!group %in% low_responses) %>% # filter out groups that should not receive feedback due to n < 6.
  pull(group)


# render files
walk(1:length(groups), function(i) {
  grupp <- groups[i]
  
  outfile <- glue("RISEfeedback_{grupp}.pptx")
  
  quarto_render(input = "AFAfeedback.qmd",
                execute_params = list("grupp" = grupp),
                output_file = outfile,
                output_format = "pptx")
})
```