1_fonts_errbars.qmd

---
title: "Using custom fonts and plotting with error bars"
author: "Michael Stadler, Charlotte Soneson"
format:
    html:
        toc: true
        toc-location: left
        embed-resources: true
        link-external-icon: true
        code-fold: true
editor_options: 
    chunk_output_type: console
---

## Summary

This document illustrates how to use `systemfonts` and `ragg` together with
`ggplot2` to:

- make a plot that uses custom fonts
- make a plot with error bars


## Prepare data

Run the following code to prepare the data used in this document:

```{r}
#| label: prepare_data
#| code-fold: false

suppressPackageStartupMessages({
    library(dplyr)
    library(tibble)
})

# built-in `mtcars` dataset (see ?mtcars)
# ... convert some columns to factors
mycars <- mtcars
mycars$cyl <- factor(mycars$cyl, levels = c("4","6","8"))
mycars$engine_shape <- factor(c("0" = "V-shaped", "1" = "straight")[as.character(mycars$vs)])
mycars$transmission <- factor(c("0" = "automatic", "1" = "manual")[as.character(mycars$am)])

# ... summarize power by cylinders
mycars_summary <- mycars |>
    group_by(cyl) |>
    summarise(hp_avg = mean(hp),
              hp_sem = sd(hp) / sqrt(length(hp)),
              .groups = "drop")

tibble(mycars)
tibble(mycars_summary)
```


## Create figure

### Load packages
```{r}
#| label: load_packages

library(ggplot2)
library(ragg)
library(systemfonts)

# remark: have to set the chunk output device to "ragg_png"
knitr::opts_chunk$set(dev = "ragg_png")
```


### Get fonts
We first need to make the fonts available to `R`. Either our desired font
is already installed on the local system, or we can download
one of the many free fonts available on the internet, for example from
[https://fonts.google.com/](https://fonts.google.com/).

The `systemfonts` package helps to list locally installed fonts that `R`
can "see" or make a downloaded font file (typically a TrueType file) available
to `R`.

```{r}
#| label: get_fonts

# list available fonts:
fonts <- systemfonts::system_fonts()
fonts
unique(fonts$family)

# download and register an additional font
# ... already registered fonts
systemfonts::registry_fonts()
# ... download Gochi Hand
download.file(url = "https://github.com/huertatipografica/gochi-hand/archive/master.zip",
              destfile = file.path(tempdir(), "master.zip"))
extracted_files <- unzip(file.path(tempdir(), "master.zip"), exdir = tempdir())
# ...  we typically want a TrueType file (extension `.ttf`)
ttf <- grep(".ttf$", extracted_files, value = TRUE)
ttf
# ... register the newly downloaded font for use in R
systemfonts::register_font(name = "Gochi Hand", plain = ttf)
# ... list the now registered fonts
systemfonts::registry_fonts()
```

### Plot
With the fonts available to `R`, we can now create plots that use custom
fonts for some or all of the text elements. Fonts are defined by
`element_text()` using the arguments `family` (see the `family` column returned
by `systemfonts` functions) and `face` (one of "plain", "italic", "bold" or
"bold.italic", see `style` column return by `systemfonts` functions).

```{r}
#| label: plot

# remember that this will only work if you have
# activated the `ragg` backend to render the plots

# create base plot
p0 <- ggplot(data = mycars, mapping = aes(x = cyl)) +
    coord_cartesian(ylim = c(0, 250)) +
    labs(x = "Number of cylinders", y = "Gross horsepower") +
    theme_bw(20) +
    theme(panel.grid = element_blank())

# ... add bars and Roman font
p0 +
    geom_col(data = mycars_summary, mapping = aes(y = hp_avg),
             fill = "gray70") +
    geom_errorbar(data = mycars_summary,
                  mapping = aes(ymin = hp_avg - hp_sem,
                                ymax = hp_avg + hp_sem),
                  width = 0.6) +
    annotate("text", x = -Inf, y = Inf, hjust = -0.1, vjust = 1.1,
             label = "Question:\nIs this font well readable?",
             family = "Z003", size = 18 / .pt) +
    theme(text = element_text(family = "Z003"))

# ... add points with errorbars and handwriting-style font
p0 +
    geom_pointrange(data = mycars_summary,
                    mapping = aes(y = hp_avg,
                                ymin = hp_avg - hp_sem,
                                ymax = hp_avg + hp_sem),
                    size = 6,
                    color = "firebrick") +
    geom_point(mapping = aes(y = hp), shape = 1) +
    annotate("text", x = -Inf, y = Inf, hjust = -0.1, vjust = 1.1,
             label = "Conclusion:\nMore cylinders give\n increased power",
             family = "Gochi Hand", size = 18 / .pt) +
    annotate("text", x = Inf, y = -Inf, hjust = 1.1, vjust = -0.5,
             label = "mean +/- sem", color = "firebrick",
             family = "Gochi Hand", size = 18 / .pt) +
    theme(text = element_text(family = "Gochi Hand"))
```


## Remarks

The page at https://r-graph-gallery.com/custom-fonts-in-R-and-ggplot2.html
describes two ways to use custom fonts with `ggplot2`:

- [`ragg`](https://ragg.r-lib.org/) is a graphics backend (it renders the plot
  to a bitmap graphics device like `png()`), and can make all system fonts
  available to be used in R. This is the approach used above. `ragg` is based
  on the fast [`AGG`](https://github.com/ghaerr/agg-2.6) library, does not
  require you to use any other packages and is also fully integrated into
  RStudio (you can select is as your default graphics device in "Tools" ->
  "Global options..." -> "General", panel "Graphics").
- [`showtext`](https://github.com/yixuan/showtext) is a package that is
  especially designed to use custom fonts with `pdf()` devices. After loading
  a font and activating `showtext`, it renders the characters as filled polygons.
  
Another useful page with more detailed information about using fonts in
`ggplot2` and `R` in general: https://yjunechoe.github.io/posts/2021-06-24-setting-up-and-debugging-custom-fonts/

## Session info

```{r}
#| label: session_info
sessioninfo::session_info()
```