03-individual_geoms.Rmd

# (PART\*) Layers {-}

# Individual Geoms

**Learning objectives**

* Discuss how geoms are the fundamental building blocks of ggplot2.
* Draw comparisons between geoms and their associated named plot.
* Explore each individual geom by reviewing their documentation.

```{r 03-individualgeoms-setup, include=FALSE}
# ggplot2 is attached with the tidyverse
library(tidyverse)

# some example data
df <- data.frame(
  c = c(3, 1, 5),
  y = c(2, 4, 6),
  label = c("a", "b", "c")
)
```

## The basics

* Each geom can be useful by itself.
* Geoms can be used in ways to construct more complex geoms.
* The geoms discussed in this chapter are two dimensional (e.g., `x` and `y`).
* All geoms understand `color` or `colour` and size aesthetics.
* Bar, tile, and polygon understand `fill`.
* The terms above are all parameters within ggplot2 functions.

## Area chart: `geom_area()`

* Draws an area plot.
  - A line plot filled to the y-axis.
  - Multiple groups are stacked.

```{r}
ggplot(diamonds, aes(x = price)) +
  geom_area(stat = "bin")
```

```{r}
ggplot(diamonds, aes(x = price, fill = cut)) +
  geom_area(stat = "bin")
```

## Bar chart: `geom_bar()`

* Makes a bar plot.

```{r}
ggplot(diamonds, aes(cut)) +
  geom_bar()
```

* What's up with `stat = "identity"`?
  - The default stat is to count values.
  - Setting this parameter leaves the data unchanged.

```{r}
# Why, though? Perhaps I want to do my own aggregation
data_diamond_count <- 
  diamonds |>
  count(cut, name = "count")

ggplot(data_diamond_count, aes(cut, count)) +
  geom_bar(stat = "identity")
```

## Line chart: `geom_line()`

* A geom that connects points from left to right.
  - `linetype` is a useful parameter.
  - Checkout the different linetypes [here](https://ggplot2.tidyverse.org/reference/aes_linetype_size_shape.html).
  - Also here `?linetype`

```{r}
ggplot(economics, aes(x = date, y = unemploy)) +
  geom_line()
```

* What's up with `geom_path()`?
  - Connects points as they appear in order of the data
  - Answer to exercise 2.

```{r}
ggplot(df, aes(c, y)) +
  geom_path()
```

```{r}
ggplot(economics, aes(unemploy / pop, psavert)) +
  geom_path()
```

## Scatterplot: `geom_point()`

```{r 03-scatter}
ggplot(mpg, aes(x = displ, y = hwy)) + 
  geom_point()
```

* The `shape` parameter is useful here.
  - interested in the different shapes? `?shape`

```{r 03-scatter-dif-shape}
ggplot(mpg, aes(x = displ, y = hwy, shape = factor(cyl))) + 
  geom_point()
```

## Polygons: `geom_polygon()`

* Draws polygons, which are filled paths.
* Useful when making maps: more in [Chapter 6](https://ggplot2-book.org/maps).

```{r}
ggplot(df, aes(c, y)) +
  geom_polygon()
```

## Histograms: `geom_histogram()`

```{r 03-hist}
ggplot(mpg, aes(hwy)) + 
  geom_histogram() 
```

## Drawing rectangles: `geom_rect()`; `geom_tile()`; `geom_raster()`

```{r}
ggplot(df, aes(c, y)) + 
  geom_tile() 
```

## Add text to a plot: `geom_text()`

* This requires the use of the `label` aesthetic, along with others

```{r}
# Filtering to simplify the example
mpg |> 
  filter(manufacturer == "ford") |>
  ggplot(aes(displ, hwy, label = model)) +
    geom_text()
```

* `position` and other parameters are also useful.

```{r}
mpg |> 
  filter(manufacturer == "ford") |>
  ggplot(aes(displ, hwy, label = model)) +
    geom_text(position = position_dodge(width = 0.2), angle = 45)
```

## Exercise solutions

### Exercise 1

* What geoms would you use to draw each of the following named plots?
  - scatterplot = `geom_point()`
  - line chart = `geom_line()`
  - histogram = `geom_histogram()`
  - bar chart = `geom_bar()` or `geom_col()`
  - pie chart = `geom_bar()` with `coord_polar()`

```{r}
ggplot(data_diamond_count, aes(cut, count)) +
  geom_col()
```

```{r}
ggplot(diamonds, aes(x = factor(1), fill = factor(cut))) +
  geom_bar(width = 1) +
  coord_polar(theta = "y")
```

### Exercise 2

* `geom_path()` connects points in order of appearance. `geom_line` connects points from left to right.

```{r 03-df, include=FALSE}
df <- data.frame(
  x = c(3, 1, 5), 
  y = c(2, 4, 6), 
  label = c("a","b","c")
)

p <- ggplot(df, aes(x, y, label = label)) + 
  labs(x = NULL, y = NULL) + # Hide axis label
  theme(plot.title = element_text(size = 12))
```

```{r 03-path}
p + geom_path()
```


* `geom_polygon()` draws polygons which are filled paths.
```{r 03-poly}
p + geom_polygon() 
```

* `geom_line()` connects points from left to right.
```{r 03-line}
p + geom_line() 
```

### Exercise 3

* What low-level geoms are used to draw geom_smooth()? 
  - `geom_smooth()` fits a smoother to data, displaying the smooth and its standard error, allowing you to see a dominant pattern within a scatterplot with a lot of "noise". The low level geom for `geom_smooth()` are `geom_path()`, `geom_area()` and `geom_point()`.

```{r 03-smooth}
ggplot(mpg, aes(displ, hwy)) + 
  geom_point() + 
  geom_smooth()
```

* What low-level geoms are used to draw geom_boxplot()? 
  - Box plots are used to summarize the distribution of a set of points using summary statistics. The low level geom for `geom_boxplot()` are `geom_rect()`, `geom_line()` and `geom_point()`. 

```{r 03-box}
ggplot(mpg, aes(drv, hwy)) + 
  geom_boxplot()
```

* What low-level geoms are used to draw geom_violin()?
  - Violin plots show a compact representation of the density of the distribution highlighting the areas where most of the points are found. The low level geom for `geom_violin()` are `geom_area()` and `geom_path()`. 

```{r 03-violin}
ggplot(mpg, aes(drv, hwy)) + 
  geom_violin()
```

## Meeting Videos

### Cohort 1

`r knitr::include_url("https://www.youtube.com/embed/XcTYBH4XbHo")`

<details>
  <summary> Meeting chat log </summary>
```
00:13:39	priyanka gagneja:	I forgot to mention that since this is relatively smaller chapter Ryan has prepared some material introducing Chapter 4 for today and he will talking about the entire chapter next week.
00:16:38	priyanka gagneja:	that's correct
00:16:42	priyanka gagneja:	that's my understanding too
00:18:38	priyanka gagneja:	what do you mean circles .. can you share a more detailed example
00:21:59	Jiwan Heo:	tibble(id = 1:10) %>% mutate(x = cos(2*pi*id/10), y = sin(2*pi*id/10)) %>% ggplot(aes(x, y)) + geom_line() + coord_equal()
00:22:05	Jiwan Heo:	vs tibble(id = 1:10) %>% mutate(x = cos(2*pi*id/10), y = sin(2*pi*id/10)) %>% ggplot(aes(x, y)) + geom_path() + coord_equal()
00:35:42	priyanka gagneja:	Thank you Ryan !!
00:38:16	priyanka gagneja:	need a min
00:52:22	Michael Haugen:	“Side rail” no pun intended
00:52:34	Ryan S:	lol
00:52:34	Michael Haugen:	sounds great
00:54:02	Ryan Metcalf:	I was going to use Derail…..no pun intended!
00:54:05	Ryan Metcalf:	Thanks you everyone!
```
</details>