A ggplot2 grammar guide

# A ggplot2 grammar guide
### Gina Reynolds, July 2019

---

A data visualization:

- is composed of geometric shapes

- that take on aesthetics

- that represent variables

- from a data set

---

We can say this another way -- a more classic definition too:

A statistical graphic 
--

maps variables of a dataset

to aesthetic properties

of geometric objects

---

# The ggplot2 grammar follows this order second order... Let's see a simple example this. (First watch the Hans Rosling clip  - )

---

```r
# specify the data for the plot
*ggplot(data = gapminder_2002)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/overview_auto_1_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
# specify the data for the plot
ggplot(data = gapminder_2002) +  
# the x position will represent
# the gdp per capita variable
* aes(x = gdpPercap)   # x position
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/overview_auto_2_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
# specify the data for the plot
ggplot(data = gapminder_2002) +  
# the x position will represent
# the gdp per capita variable
  aes(x = gdpPercap) +  # x position
# the y position will represent
# the lifeExp variable
* aes(y = lifeExp)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/overview_auto_3_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
# specify the data for the plot
ggplot(data = gapminder_2002) +  
# the x position will represent
# the gdp per capita variable
  aes(x = gdpPercap) +  # x position
# the y position will represent
# the lifeExp variable
  aes(y = lifeExp) +  
# the point geometric shape
# take on the positions according
# to the specified mapping
# i.e. the representation
* geom_point()
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/overview_auto_4_output-1.png" width="100%" />
]]

---

Let's look at each of these steps one at a time.

---

# The declarative mood: declaring data

Data is a "first class citizen" in the 'Grammar of Graphics' and `tidyverse` framework.  Therefore, the first step to creating a plot, not surprisingly, is declaring data.

Let's do it.

---

```r
ggplot(data = gapminder_2002)  # That's it
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/declare_data_1_1_output-1.png" width="100%" />
]]

---

# Pipe data into ggplot()

An alternative to the above syntax is to "pipe" the data into the ggplot function, as shown below.  The pipe operator, which is made available in R by loading the `tidyverse` package, pulls the object that preceeds it into the function.  This second option allows us to glimpse the raw data that we will use in our plot.

```r
# Option 2 
gapminder_2002 %>% # data piped into 
  ggplot() # ggplot function initiating plot
```

Let's see how this works. Then we'll move on to the second step in this grammar lesson --- the interogative mood --- aesthetic representation.

---

```r
# Option 2
*gapminder_2002   # data piped into
```
]]
.column[.content[

```
  # A tibble: 142 x 6
     country     continent  year lifeExp       pop gdpPercap
     <fct>       <fct>     <int>   <dbl>     <int>     <dbl>
   1 Afghanistan Asia       2002    42.1  25268405      727.
   2 Albania     Europe     2002    75.7   3508512     4604.
   3 Algeria     Africa     2002    71.0  31287142     5288.
   4 Angola      Africa     2002    41.0  10866106     2773.
   5 Argentina   Americas   2002    74.3  38331121     8798.
   6 Australia   Oceania    2002    80.4  19546792    30688.
   7 Austria     Europe     2002    79.0   8148312    32418.
   8 Bahrain     Asia       2002    74.8    656397    23404.
   9 Bangladesh  Asia       2002    62.0 135656790     1136.
  10 Belgium     Europe     2002    78.3  10311970    30486.
  # … with 132 more rows
```
]]
---
class: split-40
count: false

```r
# Option 2
gapminder_2002 %>%  # data piped into
* ggplot()  # ggplot function initiating plot
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/declare_data_option_2_auto_2_output-1.png" width="100%" />
]]

---

# The interogative mood:  aesthetic mapping

Aesthetics --- like position, color, and size --- represent variables which do a lot of work for us in data visualization.

Distributions in values --- which can be painstaking to digest in their "raw" form in a table of numbers or categories --- are easy to communicate when represented by position, color, size and so on.

The term 'mapping' in 'aesthetic mapping' refers to the fact that variables are 'mapped' to aesthetics -- i.e. they are represented by aesthetics.

---

# A main pool of aesthetics

Let's think about the various aesthetics that can be used to communicate distributions.

---

# A main pool of aesthetics

Here's a key set of aesthetics that can be used to communicate about distribution among categories or the spread in a variable.

Note: This figure is from Wilke's [Fundamentals of Data Visualization](https://serialmentor.com/dataviz/aesthetic-mapping.html#aesthetics-and-types-of-data).

---

# Think "ask" when you see aes()

As discussed above aes() refers to "aesthetic mapping", where "mapping" is meant in terms of representation.

What's inside aes() addresses the question, "What variable are we asking the aesthetic (color, x-position, shape, etc.) to represent?"  aes() is "asking" politely for a specific aesthetic to do us the favor of representation.

For example `aes(color = age)` can be translated to English as, "Please, color aesthetic, represent the variable `age` for me."

It is aesthetic mapping too that helps us "interogate" our data; mapped aesthetics quickly communicate distributions within variables that we may be curious about.

So it might be helpful to think "ask" when you see aes().

---

# requesting aesthetic representation

Let's see how we request an aesthetic to represent a variable from our dataset using ggplot.  We'll first look at the x position (horizontal) and y position (vertical).

---

```r
*gapminder_2002
```
]]
.column[.content[

```r
gapminder_2002 %>%  
* ggplot()
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/x_and_y_auto_2_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
gapminder_2002 %>%  
  ggplot() +  
# x position represents gdpPercap variable
* aes(x = gdpPercap)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/x_and_y_auto_3_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
gapminder_2002 %>%  
  ggplot() +  
# x position represents gdpPercap variable
  aes(x = gdpPercap) +  
# y position represents life expectancy
* aes(y = lifeExp)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/x_and_y_auto_4_output-1.png" width="100%" />
]]

---

# A complete sentance:  data + aes + geom

The aes() statement says that we want the variable `gdpPercap` represented by the x position, and the variable `lifeExp` represented by the y position.  The x and y scales give us a clue that this info is registered.

But, we don't really have any insight about our data yet.  Why?  Because aesthetics are taken on by geometric objects.  We still need to declare what geometric object will take on the aesthetics.  Let's see how, when we add a geometric object "point"; the x and y position for each row of data are taken on by that object.

---

```r
*gapminder_2002
```
]]
.column[.content[

```r
gapminder_2002 %>%  
* ggplot()
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/geom_for_aes_auto_2_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
gapminder_2002 %>%  
  ggplot() +  
* aes(x = gdpPercap)   # x position
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/geom_for_aes_auto_3_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
gapminder_2002 %>%  
  ggplot() +  
  aes(x = gdpPercap) +  # x position
* aes(y = lifeExp)   # y position
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/geom_for_aes_auto_4_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
gapminder_2002 %>%  
  ggplot() +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
# points take on x and y position
* geom_point()
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/geom_for_aes_auto_5_output-1.png" width="100%" />
]]

---

## Exploring more aesthetic mappings

Let's not get distracted by geoms (we'll come back to these).  Aesthetic mapping does so much work for us in data viz.

So, let's explore a bunch of the *aesthetic* that might represent *variables* in our data.

---

So far, we have have stated the x and y positions - these are *required* aesthetics for the "point" geometric object.

But more are *optional.*  In the next example, we'll do all of the required aesthetic mapping (x and y position) and then also use other allowable aesthetics for `geom_point()` (color, shape, size, alpha).  We'll also see that double or tripple "mapping" is allowed -- multiple aesthetics may represent the same variable.

---

```r
*ggplot(data = gapminder_2002)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_1_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
* aes(x = gdpPercap)   # x position
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_2_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
* aes(y = lifeExp)   # y position
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_3_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
* geom_point()   # above aes are required for point
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_4_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
* aes(color = continent)   # color
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_5_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
  aes(color = continent) +  # color
* aes(shape = continent)   # shape
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_6_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
  aes(color = continent) +  # color
  aes(shape = continent) +  # shape
* aes(size = pop)   # size
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_7_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
  aes(color = continent) +  # color
  aes(shape = continent) +  # shape
  aes(size = pop) +  # size
* aes(alpha = lifeExp)   # transparency
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_8_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
  aes(color = continent) +  # color
  aes(shape = continent) +  # shape
  aes(size = pop) +  # size
  aes(alpha = lifeExp) +  # transparency
* aes(color = lifeExp)  # overwriting color's representation
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point_auto_9_output-1.png" width="100%" />
]]

---

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
  aes(color = continent)   
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point1_rotate_1_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
* aes(shape = continent)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point1_rotate_2_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
* aes(size = pop)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point1_rotate_3_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
* aes(alpha = lifeExp)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point1_rotate_4_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
ggplot(data = gapminder_2002) +  
  aes(x = gdpPercap) +  # x position
  aes(y = lifeExp) +  # y position
  geom_point() +  # above aes are required for point
* aes(color = lifeExp)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_point1_rotate_5_output-1.png" width="100%" />
]]

---

# A few more aesthetics

Other aesthetics can be explored by changing the geometric object to one that allows additional aesthetics, like *fill* and *line type*; one such "geom" is geom_col(), to create a column geometry.

Based on the plot that follows, how do the fill and color aesthetic differ?

[Or I'm on the overview track, take me to the next session](#scales)

---

```r
*continent_aggregate
```
]]
.column[.content[

```
  # A tibble: 5 x 2
    continent country_count
    <fct>             <int>
  1 Africa               52
  2 Americas             25
  3 Asia                 33
  4 Europe               30
  5 Oceania               2
```
]]
---
class: split-40
count: false

```r
continent_aggregate %>%  
* ggplot()
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_area_auto_2_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
continent_aggregate %>%  
  ggplot() +  
* aes(x = continent)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_area_auto_3_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
continent_aggregate %>%  
  ggplot() +  
  aes(x = continent) +  
* aes(y = country_count)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_area_auto_4_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
continent_aggregate %>%  
  ggplot() +  
  aes(x = continent) +  
  aes(y = country_count) +  
* geom_col()
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_area_auto_5_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
continent_aggregate %>%  
  ggplot() +  
  aes(x = continent) +  
  aes(y = country_count) +  
  geom_col() +  
* aes(color = continent)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_area_auto_6_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
continent_aggregate %>%  
  ggplot() +  
  aes(x = continent) +  
  aes(y = country_count) +  
  geom_col() +  
  aes(color = continent) +  
* aes(fill = continent)   # fill color for areas
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_area_auto_7_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
continent_aggregate %>%  
  ggplot() +  
  aes(x = continent) +  
  aes(y = country_count) +  
  geom_col() +  
  aes(color = continent) +  
  aes(fill = continent) +  # fill color for areas
* aes(alpha = country_count)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_area_auto_8_output-1.png" width="100%" />
]]
---
class: split-40
count: false

```r
continent_aggregate %>%  
  ggplot() +  
  aes(x = continent) +  
  aes(y = country_count) +  
  geom_col() +  
  aes(color = continent) +  
  aes(fill = continent) +  # fill color for areas
  aes(alpha = country_count) +  
* aes(linetype = continent)
```
]]
.column[.content[
<img src="data_aes_geom_files/figure-html/aesthetics_area_auto_9_output-1.png" width="100%" />
]]

---

## Questions:

- What are *required aesthetics* vs. optional?

- What are the names of seven different aesthetics that we used above?

- What is the difference between the *fill* and *color* aesthetic?

- Look at the *help* for geom_text.  What are the *required* aesthetics?

---
name: geoms

# Nouns:  Geometric objects

So far, we have seen a few geometric objects: geom_point(), geom_col(), geom_line().

We won't cover many more here.  Why?   The celebrated language teacher, Michel Thomas, argued that building up vocabulary should not be a major focus in becoming fluent in a foreign language.  I'm of the same mind.

Geometric objects --- our "nouns" --- are abundant.  But we can use rather few of them and still work towards a profound working understanding of the ggplot2 grammar. With that work done, exchanging one geom for another, like a language learner exchanging "dog" for "canine", is a rather easy business.

But there are some grammatical things that we'll come back to, as well.

We'll see a few more geometric objects along the way as we move forward.  And, then come back to this topic squarely later on.

---

Next up: