ggplot_fonts_RLadiesFreiburg.Rmd

---
title: "How do I use custom fonts?"
author: "Julia Müller & Kyla McConnell"
date: "5 11 2021"
output: html_document
---

# Data, packages, an example graph

```{r}
options(scipen = 999)

library(showtext) # for option A
library(extrafont) # for option B
library(ggtext) # Markdown formatting in titles and labels
library(tidyverse)

survey <- read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-05-18/survey.csv') %>% 
  drop_na(gender)

head(survey)
```

Tidy Tuesday dataset contains the results of a survey on annual income of more than 24,000 people!

More info on the dataset:
https://github.com/rfordatascience/tidytuesday/blob/master/data/2021/2021-05-18/readme.md

For convenience, we'll narrow down the data to people in the US, simplify the gender column and filter out "Other" responses:
```{r}
survey_US <- survey %>% 
  filter(currency == "USD" & country %in% c("United States", "USA", "US", "U.S.", "United States of America", "Usa", "United states", "united states", "Us")) %>% 
  mutate(gender = fct_collapse(gender,
                               "Other/prefer not to answer" = c("Other or prefer not to answer", "Prefer not to answer"))) %>% 
  filter(gender != 'Other/prefer not to answer')

head(survey_US)
```

Let's create a boxplot of annual salary by gender:
```{r}
(survey_box <- survey_US %>% 
  filter(annual_salary < 1000000) %>% 
  ggplot() +
  aes(x = gender, y = annual_salary, colour = gender) +
  geom_violin() +
  geom_boxplot() +
  theme_minimal() +
  labs(x = "Gender", y = "Annual salary",
       title = "Average annual salary",
       subtitle = "N = 21091 (US respondents only)",
       caption = "Source: Tidy Tuesday - Ask a Manager Survey") +
  theme(legend.position = "none") +
  scale_color_manual(values = c("#c70039", "#ff5733", "#ffc305")))
```

...and a line graph of how salary develops over time, still split up by gender:
```{r}
(survey_dev <- survey_US %>% 
  mutate(years_of_experience_in_field = fct_relevel(years_of_experience_in_field, "1 year or less", "2 - 4 years", "5-7 years", "8 - 10 years", "11 - 20 years", "21 - 30 years", "31 - 40 years", "41 years or more"),
         years_of_experience_in_field = fct_recode(years_of_experience_in_field, "5 - 7 years" = "5-7 years")) %>% 
  group_by(gender, years_of_experience_in_field) %>% 
  summarise(avg_salary = mean(annual_salary)) %>% 
  ungroup() %>% 
  ggplot() +
  aes(x = years_of_experience_in_field, y = avg_salary,
      colour = gender, group = gender) +
  geom_point() +
  geom_line() +
  theme_minimal() +
  labs(x = "Years of professional experience in the field",
       y = "Average annual salary (in USD)",
       title = "Average annual salary",
       subtitle = "N = 21091 (US respondents only)",
       caption = "Source: Tidy Tuesday - Ask a Manager Survey") +
  scale_color_manual(values = c("#c70039", "#ff5733", "#ffc305")))
```


# Font options in default ggplot's `theme()`

## Types of texts that can be changed

![Text elements in plots](element_text.png)

theme(
    text = #all text items
    plot.title = 
    plot.subtitle = 
    plot.caption = #bottom right
    plot.tag =  #top left
    axis.title = #axis labels
    axis.title.x = 
    axis.title.y = 
    axis.text = #axis tick labels
    axis.text.x = 
    axis.text.y = 
    legend.title = 
    legend.text = 
  )


## element_text()
Most commonly used text adjustments:
```{r}
element_text(
  face = NULL, 
  color = NULL, 
  size = NULL,
  family = NULL
)
```


```{r}
survey_dev +
  theme(
    text = element_text(face = "bold") #try also italic or bold.italic
  )
```

```{r}
survey_dev +
  theme(
    axis.text = element_text(color = "purple")
  )
```


## Inbuilt font options

There are three inbuilt options in ggplot which map (point to) three fonts. Which exact fonts these are might depend on your operating system.

sans (the default) - maps to Arial
serif - maps to Times New Roman (can also call "Times")
mono - maps to Courier New

We can specify these in the `family` argument in `element_text()`:
```{r}
survey_dev +
  theme(
    plot.title = element_text(family = "serif")
  )

survey_dev +
  theme(
    plot.title = element_text(family = "mono")
  )
```


# Custom fonts

Broadly, there are two ways to use other fonts in ggplot:
- Download the font file(s) you'd like to use, then let R know where they are saved
- Download *and install* fonts, then allow R to access them. This gives R access to all fonts you have installed on your computer


## Option A: Font files not installed

### Where to find/match fonts
For example:
- fontspace.com
- fonts.google.com

dyslexic-friendly: https://t.co/grXip7myuF?amp=1

Pick one or several fonts you like and download them as a TrueType Font (file ending: .ttf). Save these files in the same folder as this script.

### Using the fonts with the {showtext} package

Use the function `font_add()` to register the font with R.
The first argument is family = the name you want to assign to the font. We'll use this in `theme()`to tell ggplot which font we'd like it to use.
The second argument is the path to the font file. If we use a Markdown document, we can simply type in the name of the file, including the ".ttf" ending.

```{r}
font_add(family = "Bungee", "Bungee-Regular.ttf")
font_add(family = "Fauna", "FaunaOne-Regular.ttf")
```

If we're using a Google font, we can instead just use the `font_add_google()` command:
```{r}
font_add_google(name = "Poiret One",   # Name of the font as it is listed on Google fonts
                family = "Poiret") # Name we'll use in theme()
```

..and one more command:
```{r}
showtext_auto() # allows fonts to show up in plots
```
You need to run this every time you add new fonts.

Now, we can specify the fonts we'd like to use in a `theme()`call. Here, we're setting all text elements to "Fauna" but the plot title to "Bungee":
```{r}
survey_box +
  theme(text = element_text(family = "Fauna"),
        plot.title = element_text(family = "Bungee"))
```

And to check that the font we directly got from Google Fonts works too:
```{r}
survey_box +
  theme(plot.title = element_text(family = "Poiret"))
```


### Try it!
a) Download a font you like, save it in the same folder as this file, and use it
b) Find a Google font you like and use it in a plot


## Option B: Font files installed

### Installing fonts

Where are fonts kept on your computer?
Windows: somewhere like C:\Windows\Fonts
Mac: Network/Library/Fonts

How do you install them?
Windows (10): Right-click -> install (can check it worked by searching in the fonts folder)

Which file types work?
Any that have the ".ttf" (preferred) or ".otf" endings


### Using the {extrafont} package

To make sure we won't run into any problems, we'll detach the showtext package before working with extrafont:
```{r}
detach("package:showtext", unload=TRUE)
```

The first time you use this package OR when you've installed new fonts, run:
```{r}
font_import() # this takes a few minutes! Make sure to select "y" and press Enter
```

Then (and every time you start a new R session), run:
```{r}
loadfonts() # for other operating systems
# OR
loadfonts(device = "win") # for Windows
```

To list the available fonts:
```{r}
fonts()
```

### Using fonts
They can now be added to plots like before:
```{r}
survey_box +
  theme(text = element_text(family = "Calibri"))
```


# Markdown formatting in titles and labels with the {ggtext} package

## Italics and bold

As we've seen, we can change the properties (bold, italic, colours, fonts, etc.) of entire elements. 
The {ggtext} package lets you use make specific words bold or italic (as opposed to the entire text in base ggplot).

First, put one asterisk around the word(s) for italics and two for bold text. Then, in `theme()`, set the text element to `element_markdown()`.
```{r}
survey_dev +
  labs(title = "*Average* annual **salary**") +
  theme(plot.title = element_markdown())
```

We need to do this for every element separately. Within `element_markdown()`, we can make use of the same options as in `element_text()`:
```{r}
survey_dev + 
  labs(title = "*Average* annual **salary**",
       subtitle = "N = 21091 *(US respondents only)*") +
  theme(plot.title = element_markdown(),
        plot.subtitle = element_markdown(size = 20, colour = "red"))
```

## Colours
To colour specific words in a text, we need CSS (Cascading Style Sheets) formatting.
As a first example, we'd like to colourcode the "US respondents only" part red.
We need to wrap that part of the text in:
`<span style = 'color:_____'> your text </span>`
```{r}
survey_dev +
  labs(subtitle = "N = 21091 <span style ='color:red'>(US respondents only)</span>") +
  theme(plot.subtitle = element_markdown())
```

We've changed our minds, would actually prefer blue and have found the hex code #100257. This works the same way:
```{r}
survey_dev +
  labs(subtitle = "N = 21091 <span style ='color:#100257'>(US respondents only)</span>") +
  theme(plot.subtitle = element_markdown())
```

For a more useful example, we'll explain what the colours in the plot mean by colourcoding words in the title. This way, we can get rid of the separate legend:
```{r}
survey_dev +
  labs(title = "Average annual salary for 
       <span style ='color:#c70039'>men</span>,
       <span style ='color:#ffc305'>women</span> and
       <span style ='color:#ff5733'>non-binary people</span>") +
  theme(plot.title = element_markdown(),
        legend.position = "null")
```

Similarly, we can change the fonts and font sizes by adding arguments to the CSS. These arguments need to be separated by a semicolon:
`<span style = 'color:_____;font-family:_____;font-size:____' your text </span>`
```{r}
survey_dev +
  labs(title = "Average annual salary for 
       <span style ='color:#c70039;font-family:Bungee'>men</span>,
       <span style ='color:#ffc305;font-size:24pt'>women</span> and
       <span style ='color:#ff5733;'>non-binary people</span>") +
  theme(plot.title = element_markdown(),
        legend.position = "null")
```

Add line breaks with <br>
```{r}
survey_dev +
  labs(title = "Average annual salary for <br>
       <span style ='color:#c70039;font-family:Bungee'>men</span>,
       <span style ='color:#ffc305;font-size:24pt'>women</span> and
       <span style ='color:#ff5733;'>non-binary people</span>") +
  theme(plot.title = element_markdown(),
        legend.position = "null")
```

The markdown syntax for italics and bold text also still work within <span>
```{r}
survey_dev +
  labs(title = "Average annual salary for <br>
       <span style ='color:#c70039;font-family:Bungee'>men</span>,
       <span style ='color:#ffc305;font-size:24pt'>women</span> and
       <span style ='color:#ff5733;'>**non-binary people**</span>") +
  theme(plot.title = element_markdown(),
        legend.position = "null")
```

...and we can change fonts for all texts in `theme()` still and it'll apply to `element_markdown()`-elements too:
```{r}
survey_dev +
  labs(title = "*Average* annual **salary** for <br>
       <span style ='color:#c70039;font-family:Bungee'>men</span>,
       <span style ='color:#ffc305;font-size:24pt'>women</span> and
       <span style ='color:#ff5733;'>**non-binary people**</span>") +
  theme(text = element_text(family = "serif"),
        plot.title = element_markdown(),
        legend.position = "null")
```

## Text boxes
We can also draw a textbox around text with `element_textbox_simple()` in `theme()`:
```{r}
survey_dev +
  theme(plot.title = element_textbox_simple(
        size = 18, # font size
        face = "bold",
        family = "Bungee",
        colour = "red",
        linetype = 1, # turn on border
        box.color = "darkgray", # border color (hex codes possible)
        fill = "gray", # background fill color (hex codes possible)
        padding = margin(5, 5, 5, 5), # padding around text inside the box
        margin = margin(10, 0, 10, 0))) # margin around text box
```

We can combine textboxes and the markdown formatting we've been using:
```{r}
survey_dev +
  labs(title =  "Average annual salary for <span style = 'color:#c70039;'>men</span>, <span style = 'color:color:#ffc305;'>women</span> and
       <span style ='color:#ff5733'>non-binary people</span>") +
  theme(plot.title = element_textbox_simple(
        #size = 18,
        linetype = 1, # turn on border
        box.color = "darkgray", # border color (hex codes possible)
        fill = "lightgray", # background fill color (hex codes possible)
        padding = margin(5, 5, 5, 5), # padding around text inside the box
        margin = margin(0, 0, 10, 0)),  # margin around text box
        legend.position = "null")
```