Plotting

class: center, middle, inverse, title-slide

# Plotting
## with ggplot2
### Basel R Bootcamp <a href='https://therbootcamp.github.io'>www.therbootcamp.com</a> <a href='https://twitter.com/therbootcamp'>@therbootcamp</a>
### July 2018

---

layout: true

<div class="my-footer">
<a href="https://therbootcamp.github.io/">BaselRBootcamp, July 2018</a>
&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;
&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;
&emsp;&emsp;&emsp;&emsp;&emsp;
&emsp;&emsp;&emsp;&emsp;&emsp;
<a href="https://therbootcamp.github.io/">www.therbootcamp.com</a>
</div>

---

.pull-left4[
 
> ### As good as R is for statistics, it's as good if not better for data visualisation
> ### Nathaniel D. Phillips
]

.pull-right6[
 
<img src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/ggplotgallery.png" width="100%" style="display: block; margin: auto;" />

]

---

.pull-left45[

# Base R Plotting

The <high>classic framework</high> of plotting.

Contains separate <high>function for each 'type'</high> of plot.

E.g. `barplot()` for a bar plot, `boxplot()` for a box plot, and `plot()` for a scatterplot.

```r
# Histogram in base R
hist(x = baselers$age,
     xlab = "Age",
     ylab = "Frequency",
     main = "Baselers Age")
```

]

.pull-right5[
 
<img src="Plotting_files/figure-html/unnamed-chunk-5-1.png" style="display: block; margin: auto;" />

]

---

.pull-left45[

# Base R Plotting

The <high>classic framework</high> of plotting.

Contains separate <high>function for each 'type'</high> of plot.

E.g. `barplot()` for a bar plot, `boxplot()` for a box plot, and `plot()` for a scatterplot.

```r
# Boxplot in base R
boxplot(formula = height ~ sex,
        data = baselers,
        xlab = "Sex",
        ylab = "Height",
        main = "Box plot")
```

]

.pull-right45[
 
<img src="Plotting_files/figure-html/unnamed-chunk-7-1.png" style="display: block; margin: auto;" />

]

---

.pull-left45[

# Base R Plotting

The <high>classic framework</high> of plotting.

Contains separate <high>function for each 'type'</high> of plot.

E.g. `barplot()` for a bar plot, `boxplot()` for a box plot, and `plot()` for a scatterplot.

```r
# Scatterplot in base R
plot(x = baselers$height,
     y = baselers$income,
     xlab = "Height",
     ylab = "Income",
     main = "Scatterplot")
```

]

.pull-right45[
 
<img src="Plotting_files/figure-html/unnamed-chunk-9-1.png" style="display: block; margin: auto;" />

]

---

# Problems with Base R plotting

.pull-left35[

- Default plots look pretty <high>outdated</high>.

- Plots can quickly require a <high>LOT of code</high>.

- Can't store plots as <high>objects</high> to reference and update later

<high>Solution: `ggplot2`</high>

]

.pull-right55[

This plot would take <high>a lot of code in Base R</high> but <high>just 10 lines of code</high>, 5 of which controlling the labels, in `ggplot2`.

]

---

# Grammar of Graphics in `ggplot2`

.pull-left45[

The Grammar of graphics breaks down plots into several key pieces:

| Aesthetics| Description|
|:------|:----|
| Data| What dataframe contains the data?|
| axes| What does the x-axis, y-axis, color (etc) represent?|
| color| What does color represent? |
| size | What does size represent? |
| geometries| What kind of geometric object do you want to plot?|
| facets| Should there be groups of plots?|

]

.pull-right45[

]

---

# Our goal: Creating this plot

.pull-left45[

- Use the `mpg` tibble

<high>Aesthetics</high>

- Engine displacement (`disp`) on the x axis
- Highway miles per gallon (`hwy`) on the y-axis
- Color plotting elements by the `class` of car

<high>Geometric objects</high>

- Show data as points
- Add a regression line

<high>Labels and themes</high>

- Add plotting labels
- Use a black and white plotting theme

]

.pull-right5[
 
<img src="Plotting_files/figure-html/unnamed-chunk-13-1.png" style="display: block; margin: auto;" />

]

---

# `ggplot`

.pull-left45[

To <high>create a ggplot2 object</high>, use the `ggplot()` function

`ggplot()` has two main arguments:

- `data` - A data frame (aka `tibble)
- `mapping` - A call to `aes()`

]

.pull-right45[

```r
ggplot(data = mpg)
```

]

---

# `ggplot`

.pull-left4[

An <high>aesthetic mapping</high> is a visual property of the objects in your plot.

Use `aes()` to assign columns in your dataframe to properties in your plot.

Common aesthetics are...

| aesthetics| Description|
|:------|:----|
|     `x`, `y`|    Data mapped to coordinates|
|     `color`, `fill`|    Border and fill colors|
|     `alpha`| Transparency|
|    `size`|  Size|
|    `shape`|  Shape|

]

.pull-right5[

```r
ggplot(data = mpg,
       mapping = aes(x = displ, y = hwy))
```

]

---

# Adding elements to plots with '+'

.pull-left45[

Once you have specified the `data` argument, and global aesthetics with `mapping = aes()`, <high>add additional elements to the plot with `+`</high>.

The `+` operator works just like the pipe `%>%` in `dplyr`. <high>It just means "and then..."</high>

```r
ggplot(data = mpg,
       mapping = aes(x = displ,
                     y = hwy)) + #and then
       geom_point() 
```

]

.pull-right5[

]

---

# Geometric objects (`geom`)

.pull-left4[

A <high>`geom`</high> is a geometric object in a plot that represents data

To add a geom to a plot, just include ` + geom_X()` where X is the type of geom.

Common geoms are...

| geom| output|
|:------|:----|
|     `geom_point()`|    Points|
|     `geom_bar()`|    Bar|
|     `geom_boxplot()`| Boxplot
|     `geom_count()`| Points with size reflecting frequency|
|     `geom_smooth()`|  Smoothed line|

]

.pull-right5[

]

---

.pull-left45[

## `geom_boxplot`

```r
ggplot(data = mpg,
  mapping = aes(x = class,
                y = hwy)) + 
  geom_boxplot()
```

]

.pull-right45[

## `geom_violin`

```r
ggplot(data = mpg,
  mapping = aes(x = class,
                y = hwy)) + 
  geom_violin()
```

]

---

.pull-left45[

## `geom_bar`

```r
ggplot(data = mpg,
  mapping = aes(x = class)) + 
  geom_bar()
```

]

.pull-right45[

## `geom_count`

```r
ggplot(data = mpg,
  mapping = aes(x = displ,
                y = hwy)) + 
  geom_count()
```

]

---

# `aes`

.pull-left45[

`color` geoms according to a variable.

```r
ggplot(data = mpg,
       mapping = aes(x = displ,
                     y = hwy,
                     color = class)) + 
  geom_point()
```

`mpg`

| displ| hwy|class | year|
|-----:|---:|:----------|----:|
<<<<<<< HEAD
| 6.2| 26|2seater | 2008|
| 5.4| 17|suv | 1999|
| 1.6| 33|subcompact | 1999|
| 2.0| 28|subcompact | 2008|
| 3.0| 26|midsize | 1999|
=======
| 4.0| 20|pickup | 2008|
| 2.2| 27|compact | 1999|
| 5.2| 17|pickup | 1999|
| 4.0| 24|subcompact | 2008|
| 4.6| 21|subcompact | 1999|
>>>>>>> c27f4a99084727da7380c82dba27ad76198715ed

]

.pull-right5[

]

---

# What's next?

.pull-left45[

]

.pull-right45[

]

---

# `geom_smooth`

.pull-left45[

`geom_smooth()` adds a smoothed (average) line.

Change how the line is created with `method` (e.g.; method = 'lm').

Color the line with `col`.

```r
ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy,
                     col = class)) +
  geom_point() + 
  geom_smooth(col = "blue")
```

]

.pull-right45[

]

---

# `geom_smooth`

.pull-left45[

`geom_smooth()` adds a smoothed (average) line.

Change how the line is created with `method` (e.g.; method = 'lm')

Color the line with `col`

```r
ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy,
                     col = class)) +
  geom_point() + 
  geom_smooth(col = "blue",
              method = "lm")
```

]

.pull-right45[

]

---

# Overriding aesthetics

.pull-left45[

If you add additional plotting aesthetics, they will <high>override</high> the general plotting aesthetics.

This is what happens, when you don't override...

```r
ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy,
                     col = class)) +
  geom_point() + 
  geom_smooth() # no overriding
```

]

.pull-right45[

]

---

# What's next?

.pull-left5[

]

.pull-right45[

]

---

# `labs`

.pull-left45[

You can add labels to a plot with the `labs()` function

`labs()` arguments are ...

- `title` - Main title
- `subtitle` - Subtitle
- `caption` - Caption below

```r
ggplot(...) +
  labs(x = "Engine Displ...",
       y = "Highway miles...",
       title = "MPG data",
       subtitle = "Cars with ...",
       caption = "Source...")
```

]

.pull-right45[

]

---

# What's next?

.pull-left5[

]

.pull-right45[

]

---

# Themes with `theme_XX`

.pull-left45[

A plotting <high>theme</high> controls many aspects of its <high>overall look</high>, from the background, to the grid lines, to the label font to the spacing between plot labels and the plotting space.

Themes built into `ggplot2`

`theme_bw()` 
`theme_minimal()` 
`theme_classic()` 
`theme_light()` 
`theme_void()`

Themes from the `ggthemes` package

`theme_excel()` 
`theme_economist()` 
*And many more!!*

]

.pull-right45[

```r
ggplot(...) + 
  theme_gray() # The Default theme
```

]

---

# Themes with `theme_XX`

.pull-left45[

Themes built into `ggplot2`

`theme_bw()` 
`theme_minimal()` 
`theme_classic()` 
`theme_light()` 
`theme_void()`

Themes from the `ggthemes` package

`theme_excel()` 
`theme_economist()` 
*And many more!!*

]

.pull-right45[

```r
ggplot(...) + 
  theme_light()
```

]

---

# Themes with `theme_XX`

.pull-left45[

Themes built into `ggplot2`

`theme_bw()` 
`theme_minimal()` 
`theme_classic()` 
`theme_light()` 
`theme_void()`

Themes from the `ggthemes` package

`theme_excel()` 
`theme_economist()` 
*And many more!!*

]

.pull-right45[

```r
ggplot(...) + 
  theme_void()
```

]

---

# Themes with `theme_XX`

.pull-left45[

Themes built into `ggplot2`

`theme_bw()` 
`theme_minimal()` 
`theme_classic()` 
`theme_light()` 
`theme_void()`

Themes from the `ggthemes` package

`theme_excel()` 
`theme_economist()` 
*And many more!!*

]

.pull-right45[

```r
library(ggthemes) # Contains many themes!

ggplot(...) + 
  theme_excel()
```

]

---

# Themes with `theme_XX`

.pull-left45[

Themes built into `ggplot2`:

`theme_bw()` 
`theme_minimal()` 
`theme_classic()` 
`theme_light()` 
`theme_void()`

Themes from the `ggthemes` package

`theme_excel()` 
`theme_economist()` 
*And many more!!*

]

.pull-right45[

```r
library(ggthemes) # Contains many themes!

ggplot(...) + 
  theme_economist()
```

]

---

## Final result!

```r
ggplot(data = mpg, 
       mapping = aes(x = displ, y = hwy, col = class)) +
  geom_point() + 
  geom_smooth(col = "blue", method = "lm")+
  labs(x = "Engine Displacement in Liters",
       y = "Highway miles per gallon",
       title = "MPG data",
       subtitle = "Cars with higher engine displacement tend to have lower highway mpg",
       caption = "Source: mpg data in ggplot2") + 
  theme_bw()
```

---

# `facet_wrap`

.pull-left4[

Faceting = Create different plots for different groups

To facet plots, use `facet_wrap()`

```r
# Without faceting
ggplot(data = mpg, 
       mapping = aes(x = displ, 
                     y = hwy)) + 
  geom_point()
```

]

.pull-right55[

]

---

# `facet_wrap`

.pull-left4[

Faceting = Create different plots for different groups

To facet plots, use `facet_wrap()`

```r
# With faceting
ggplot(data = mpg, 
       mapping = aes(x = displ, 
                     y = hwy)) + 
  geom_point()  +
  facet_wrap(~ class)  # Tilde first
```

]

.pull-right55[

]

---

# Assigning a ggplot to an object

.pull-left4[

1) ggplot returns an object of the class "gg". 
2) You can assign the result of `ggplot` to an object. 
3) Evaluating the object will show the plot. 
4) You can even edit existing `ggplot` objects.

```r
# Create myplot
myplot <- ggplot(data = mpg,
 aes(x = displ,
 y = hwy)) + 
 geom_point() + 
 theme_bw()
class(myplot)
```

```
[1] "gg"     "ggplot"
```

]

.pull-right5[

```r
myplot # Evaluate myplot
```

]

---

# Assigning a ggplot to an object

.pull-left4[

```r
# Create myplot
myplot <- ggplot(data = mpg,
 aes(x = displ,
 y = hwy)) + 
 geom_point() + 
 theme_bw()
class(myplot)
```

```
[1] "gg"     "ggplot"
```

]

.pull-right5[

```r
myplot + 
  geom_smooth() # add geom
```

]

---

# `ggsave()`

.pull-left5[

To save plots to a file (e.g.; .jpg, .pdf, .png), use the `ggsave()` function.

`ggsave()` main arguments are...

|Argument| Definition|
|:-------|:----------|
|`filename`|File name|
|`plot`|Plotting object|
|`device`|File type (e.g.; "pdf", "jpeg", "png")|
|`path`|File path to save plot|
|`width`|Plot width (inches)|
|`height`|Plot height (inches)|

]

.pull-right45[

Save ggplot object called myplot to a pdf file...

```r
# Create myplot object
myplot <- ggplot(...)

# Create "myplot.pdf", from myplot
ggsave(filename = "myplot.pdf",
       plot = myplot,
       device = "pdf",
       path = "figures",
       width = 6, 
       height = 4)
```

]

---

.pull-left45[

# `ggplot2` extensions

ggplot2 has inspired <high>hundreds</high> of extension packages that make it even easier to make domain specific plots

Check out: [`ggplot2` Extension Gallery](http://www.ggplot2-exts.org/gallery/)

```r
# Create a survival plot using ggsurvplot
library(survminer)
library(survival)

fit <- survfit(Surv(time, status) ~ sex, 
 data = lung)
ggsurvplot(fit, data = lung)
```

]

.pull-right5[

]

---

# Practical

<a href="https://therbootcamp.github.io/BaselRBootcamp_2018July/_sessions/Plotting/Plotting.html">Link to practical</a>