Graphics with ggplot2 📊

class: center, middle, inverse, title-slide

# Graphics with ggplot2 📊
## NHS R Conference 2020
### John McIntyre

---

layout:true
layout: true

<div class="jr-header">
        <img class="logo" src="assets/white_logo_full.png"/>
        <span class="social"><table><tr><td><img src="assets/twitter.gif"/></td><td> @jumping_uk</td></tr></table></span>
      </div>
      
<div class="jr-footer"><span>&copy; 2021 Jumping Rivers (jumpingrivers.com)</span><div>http://bit.ly/nhs-ggplot2</div></div>

---

# Introduction to {ggplot2}

* http://bit.ly/nhs-ggplot2

.center[
<img src="assets/graphics.jpg" width = 400></img>
]

---

layout:true
layout: true

---
background-image: url(assets/white_logo.png)
class: center, middle, inverse

# Who am I?

---

layout: true
layout: true

---

# Jumping Rivers

.pull-left[
<img src="assets/robot.jpg"></img>
]

.pull-right[

* On-site training

* R and python consultancy

* Dashboard creation

* Code review

* Questionnaire design

* R Package development

* Predictive analytics

* Grant applications
]

---

# Our clients

---

# Introduction

*  _Many_ different ways to make graphs in R

* {ggplot2} started in 2005 and follows the "Grammar of Graphics"

* Many companies have adopted {ggplot2} for graphics, including the [BBC](https://bbc.github.io/rcookbook/) and the [FT](https://twitter.com/jburnmurdoch)

* Think about graphics in terms of [layers](https://evamaerey.github.io/ggplot_flipbook/ggplot_flipbook_xaringan.html)

.center[
<img src="assets/canvas.jpg" width=400></img>
]

---

# The basic plot object

* Load the package with

```r
library("ggplot2")
```

* Create an initial ggplot object, using `ggplot()`

* This function has two arguments:

* __data__: this must be a data frame (or tibble)

* an aesthetic __mapping__: this tells {ggplot2} how to map data to the graphical elements

---

# Setting up the plot

```r
movies = readRDS("data/movies.rds")
```

* The function `aes()` maps our data to the graph

* Here, duration is mapped to the x-axis

```r
g = ggplot(data = movies,
           mapping = aes(x = duration, y = rating))
```

Notice we can store graphs in variables.

---

# Scatter plots

* To add information we need to add a `geom`.
* Can have multiple `geoms` on a graph

```r
h = ggplot(movies, aes(x = duration, y = rating))
h + geom_point()
```

---

# Bar charts

* Great for displaying qualitative data.
* Length of the bar represents the frequency

```r
ggplot(movies, aes(x = classification)) + 
  geom_bar()
```

---
class: center, middle, inverse
background-image: url(assets/white_logo.png)
background-size: cover

# Quick Quiz!

---

# Spot the mistakes

Sarah is trying to recreate the following plot, but her code doesn't work

Spot three errors in her code and think about how you would fix them.

```r
ggplot(data = movies, x = votes, y = rating) %>%
  geom_scatter()
```

---

# Histograms

* Good for plotting continuous variables.
* Data is split up into intervals called `classes`.
* Area of the columns represent the frequencies in the classes.

```r
ggplot(movies, aes(x = duration)) + geom_histogram(binwidth = 10)
```

---

# Box and whisker plots

* Central bar represents the median

* Top and bottom of the box represents the lower and upper quartiles

```r
boxplot = ggplot(movies, aes(x = classification, y = rating)) + 
   geom_boxplot()
```

---

# Titles and legends

```r
boxplot + labs(x = "Movie classification", 
         y = "User rating (1-10)", 
         title = "Movie ratings conditional on classification", 
         subtitle = "Data collected from the IMDB", 
         caption = "These box plots represent etc etc")
```

---

# Saving graphs to files

* So how do you save all your fantastic plots?

* RStudio has the *export* button

* It's also possible to export using code (better for reproducibility)

---
layout: true
layout: true

<div class="jr-header-inverse">
        <img class="logo" src="assets/white_logo_full.png"/>
        <span class="social"><table><tr><td><img src="assets/twitter.gif"/></td><td> @jumping_uk</td></tr></table></span>
      </div>
      
<div class="jr-footer-inverse"><span>&copy; 2021 Jumping Rivers (jumpingrivers.com)</span><div>http://bit.ly/nhs-ggplot2</div></div>

---

class: center, middle, inverse
background-image: url(assets/white_logo.png)
background-size: cover

# Practical

---
layout: true
layout: true

# Aesthetics

There are many different types of aesthetics

* Colour/fill 🎨

* Shape 🔺

* Size 🐘

* Linetype ✏️

* Alpha 👻

---

# Colours

```r
ggplot(movies, aes(x = duration, y = rating)) + 
*    geom_point(aes(colour = duration)) 
```

---

# Colours

```r
ggplot(movies, aes(x = duration, y = rating)) + 
*    geom_point(aes(colour = classification)) 
```

---

# Colours

```r
ggplot(movies, aes(x = duration, y = rating)) + 
*    geom_point(aes(colour = "Blue")) 
```

---

# Colours

```r
ggplot(movies, aes(x = duration, y = rating)) + 
*    geom_point(colour = "Blue") 
```

---

# Remember! ⚠️

## Inside aes():

Using columns of the data to style. E.g. `aes(colour = classification)`

## Outside aes():

Fixed styling. E.g. `colour = "blue"`

---

# Shapes

---

# Change shape for *all* points

```r
ggplot(movies, aes(x = duration, y = rating)) + 
*  geom_point(shape = 9, size = 3)
```

---

# Change shape based on the data

```r
ggplot(movies, aes(x = duration, y = rating)) + 
*  geom_point(aes(shape = classification), size = 3)
```

---
class: center, middle, inverse
background-image: url(assets/white_logo.png)
background-size: cover

# Quick Quiz!

---

# Fill in the blanks

```r
ggplot(movies, aes(x = year, y = rating)) + 
  geom_point(aes(___ = classification,
                 size = ___),
             ___ = 0.3) 
```

---

# Facets - mini plots!

```r
ggplot(movies, aes(x = duration)) + 
  geom_histogram(binwidth = 10) +
  facet_wrap(~classification)
```

---

# Themes

```
#> Warning: `guides(<scale> = FALSE)` is deprecated. Please use `guides(<scale> =
#> "none")` instead.
```
 * Quick and easy way to add style to your plots.
 * Eight in-built themes with {ggplot2}

```r
ggplot(bond, aes(Kills, Alcohol_Units)) + 
  geom_point(aes(colour = Actor)) +
  theme_bw() 
```

---

# Other themes: {ggthemes}

```r
library("ggthemes")
g + theme_excel() + scale_colour_excel()
g + theme_minimal() + scale_color_ptol()
g + theme_hc() + scale_colour_hc()
```
<img src="slides_files/figure-html/unnamed-chunk-24-1.svg" width="70%" style="display: block; margin: auto;" />

---

# Other themes: {hrbrthemes}

See https://github.com/hrbrmstr/hrbrthemes

---

# Other themes: {tvthemes}

You can even style your plot based on your favourite TV show with [{tvthemes}](https://ryo-n7.github.io/2019-05-16-introducing-tvthemes-package/)

---
class: center, middle, inverse
background-image: url(assets/white_logo.png)
background-size: cover

# Practical

---

# Top troubleshooting tips 💁

* {ggplot2} is the 📦,  `ggplot()` is the function

* `colour = "blue"` vs `aes(colour = gender)`

* Round brackets! Geoms are functions. E.g. `geom_point()`

* Remember the `+`

* Are you plotting something sensible (is your data categorical?)

* Try manipulating your data **before** plotting it

* 📊 Just `x` - `geom_bar()`, `x` and `y` - `geom_col()`

---

# Links & resources 📚

* [R4DS book chapter 3](https://r4ds.had.co.nz/data-visualisation.html)
* [R graphics cookbook](http://www.cookbook-r.com/Graphs/)
* [RStudio {ggplot2} cheatsheet](https://rstudio.com/resources/cheatsheets/)
* [#TidyTuesday](https://github.com/rfordatascience/tidytuesday)
* [Our training courses](https://www.jumpingrivers.com/training/all-courses/)

Share any cool plots you make on twitter!

@jumping_uk #NHSR #NHSRconf2021