gganimate (with a spooky twist)

.title[
# <img src="figures/bat-cartoon.png" /> gganimate (with a spooky twist)
]
.author[
### Katherine Goode 
]
.date[
### ISU Graphics Group - October 31, 2019 
]

---

<style>

.remark-slide-content {
  background-color: #FFFFFF;
  border-top: 80px solid #F9C389;
  font-size: 20px;
  font-weight: 300;
  line-height: 1.5;
  padding: 1em 2em 1em 2em
}

.inverse {
  background-color: #696767;
  border-top: 80px solid #696767;
  text-shadow: none;
  background-image: url(https://github.com/goodekat/presentations/blob/master/2019-isugg-gganimate-spooky/figures/spider.png?raw=true);
	background-position: 50% 75%;
  background-size: 150px;
}

.your-turn{
  background-color: #8C7E95;
  border-top: 80px solid #F9C389;
  text-shadow: none;
  background-image: url(https://github.com/goodekat/presentations/blob/master/2019-isugg-gganimate-spooky/figures/spider.png?raw=true);
	background-position: 95% 90%;
  background-size: 75px;
}

.title-slide {
  background-color: #F9C389;
  border-top: 80px solid #F9C389;
  background-image: none;
}

.title-slide > h1  {
  color: #111111;
  font-size: 40px;
  text-shadow: none;
  font-weight: 400;
  text-align: left;
  margin-left: 15px;
  padding-top: 80px;
}
.title-slide > h2  {
  margin-top: -25px;
  padding-bottom: -20px;
  color: #111111;
  text-shadow: none;
  font-weight: 300;
  font-size: 35px;
  text-align: left;
  margin-left: 15px;
}
.title-slide > h3  {
  color: #111111;
  text-shadow: none;
  font-weight: 300;
  font-size: 25px;
  text-align: left;
  margin-left: 15px;
  margin-bottom: -30px;
}

</style>

# Overview

2. Transition functions

3. View functions

4. Shadow functions

5. Other
]

.pull-right[
<img src="figures/cat.jpg" width="807" height="500px" style="display: block; margin: auto;" />
]

---

# Introduction

---

# What is gganimate?

- An extension to ggplot2 for creating animated graphics in R

- Originally written by David Robinson

- Updated by Thomas Lin Pedersen in 2018/2019

- Code from a year(ish) ago will probably not work anymore

- Uses the R package [tweenr](https://github.com/thomasp85/tweenr) (also by Lin Pedersen)

- Pedersen describes it as...

> "a theoretical backbone of how to describe animation, and it's also an implementation of set theoretical backbones."

---

# gapminder Example

---

# How does it work?

From the documentation...

> gganimate provides "a range of new grammar classes that can be added to the plot object in order to customize how it should change with time."

---

# Set up

Packages to load if you are following along

```r
# Installs gganimate
install.packages("gganimate")

# Loads the library
library(gganimate)

# Other libraries that will be used
library(ggplot2)
library(tidyverse)
```

---

# Bat Data

Study recorded the flying trajectories of bats (individual and paired)

- **Paper**: Giuggioli L, McKetterick TJ, Holderied M (2015) Delayed response and biosonar perception explain movement coordination in trawling bats. PLOS Computational Biology. [doi:10.1371/journal.pcbi.1004089.t001](https://journals.plos.org/ploscompbiol/article/figure?id=10.1371/journal.pcbi.1004089.t001)

- **Original Data**: Holderied M, Giuggioli L, McKetterick TJ (2015) Data from: Delayed response and biosonar perception explain movement coordination in trawling bats. Movebank Data Repository. [doi:10.5441/001/1.62h1f7k9](https://www.datarepository.movebank.org/handle/10255/move.421)

- **My Subset of the Data**:
    - extracted the first 5 individual bat trajectories
    - selected 4 variables
        - `id`: bat id
        - `time`: time of observed location (in milliseconds)
        - `latitude`: latitude at the time of observation
        - `longitude`: longitude at the time of observation

---

# Bat Data

**How to access my subset of the data:**

```r
# Link where the data are stored
link <- "https://raw.githubusercontent.com/goodekat/presentations/master/2019-isugg-gganimate-spooky/bat-data/bats-subset.csv"
```

```r
# Access the data and convert id to a factor
bats <- read.csv(link) %>% 
 mutate(id = factor(id))
```

```r
# Data structure
str(bats)
```

```
## 'data.frame':	1743 obs. of  4 variables:
##  $ id       : Factor w/ 5 levels "1","2","3","4",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ time     : num  0 0.019 0.039 0.059 0.079 0.099 0.119 0.139 0.16 0.18 ...
##  $ latitude : num  51.4 51.4 51.4 51.4 51.4 ...
##  $ longitude: num  -2.66 -2.66 -2.66 -2.66 -2.66 ...
```

---

# Bat Data

Plot of the 5 bat trajectories

---

# Basics for creating a gganimation

**Step 1**: Start with a ggplot data and mapping

```r
*ggplot(bats %>% filter(id == 1),
* aes(x = longitude,
* y = latitude,
* color = time))
```
]
.right-plot[
<img src="slides_files/figure-html/output1-1.png" style="display: block; margin: auto;" />
]

---

# Basics for creating a gganimation

**Step 2**: Add layers using geoms to create static

```r
ggplot(bats %>% filter(id == 1), 
 aes(x = longitude,
 y = latitude,
 color = time)) + 
* geom_point()
```
]
.right-plot[
<img src="slides_files/figure-html/output2-1.png" style="display: block; margin: auto;" />
]

---

# Basics for creating a gganimation

**Step 3**: Add formatting options

```r
ggplot(bats %>% filter(id == 1), 
 aes(x = longitude,
 y = latitude,
 color = time)) + 
 geom_point() +
* scale_color_gradient2(
* midpoint = 6,
* low = "orange",
* mid = "purple",
* high = "black")
```
]
.right-plot[
<img src="slides_files/figure-html/output3-1.png" style="display: block; margin: auto;" />
]

---

# Basics for creating a gganimation

**Step 4**: Add gganimate options

```r
ggplot(bats %>% filter(id == 1), 
 aes(x = longitude,
 y = latitude,
 color = time)) + 
 geom_point() +
 scale_color_gradient2(
 midpoint = 6,
 low = "orange",
 mid = "purple",
 high = "black") +
* transition_reveal(time)
```
]
.right-plot[
<img src="slides_files/figure-html/output4-1.gif" style="display: block; margin: auto;" />
]

---

# Rendering process

---

# Transitions

---

# Transition: States

`transition_states`: Transition between several *distinct stages* of the data

Options:

- `states`: The unquoted name of the column holding the state levels in the data.

- `transition_length`: The relative length of the transition. Will be recycled to match the number of states in the data

- `state_length`:	The relative length of the pause at the states. Will be recycled to match the number of states in the data

- `wrap = TRUE`: Should the animation wrap-around? If TRUE the last state will be transitioned into the first.

---

# Transition: States

Static plot of the 5 trajectories colored by time

```r
bats %>%
 ggplot(aes(x = longitude, 
 y = latitude, 
 color = time)) +
 geom_point() + 
 scale_color_gradient2(
 midpoint = 6, 
 low = "orange", 
 mid = "purple",
 high = "black")
```
]
.right-plot[
<img src="slides_files/figure-html/statesout1-1.png" style="display: block; margin: auto;" />
]

---

# Transition: States

Add `transition_states` with `states = id`

Something is a bit strange...

```r
bats %>%
 ggplot(aes(x = longitude, 
 y = latitude, 
 color = time)) +
 geom_point() + 
 scale_color_gradient2(
 midpoint = 6, 
 low = "orange", 
 mid = "purple",
 high = "black") + 
* transition_states(states = id)
```
]
.right-plot[
<img src="slides_files/figure-html/statesout2-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: States

From the documentation on the importance of specifying the `group`:

> The group aesthetic, if not set, will be calculated from the interaction of all discrete aesthetics in the layer (excluding label)

> It is often better to set it explicitly when animating, to make sure your data is interpreted in the right way.

> Unmatched rows will appear/disappear, potentially using an enter or exit function.

---

# Transition: States

Add `group = id`

```r
bats %>%
 ggplot(aes(x = longitude, 
 y = latitude, 
 color = time, 
* group = id)) +
 geom_point() + 
 scale_color_gradient2(
 midpoint = 6, 
 low = "orange", 
 mid = "purple",
 high = "black") + 
 transition_states(states = id)
```
]
.right-plot[
<img src="slides_files/figure-html/statesout3-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: States

Add a label to describe the state and change the transition length and time

```r
bats %>%
 ggplot(aes(x = longitude, 
 y = latitude, 
 color = time,
 group = id)) +
 geom_point() + 
 scale_color_gradient2(
 midpoint = 6, 
 low = "orange", 
 mid = "purple",
 high = "black") + 
* labs(title =
* "Bat {previous_state}") +
 transition_states(
 states = id, 
* transition_length = 3,
* state_length = 3)
```
]
.right-plot[
<img src="slides_files/figure-html/statesout4-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: Time

`transition_time`: Transition through distinct states in *time*

Options:

- `time`: An unquoted expression giving the time, and thus state membership, of each observation.

- `range`: The time range to animate. If NULL it will be set to the range of time

---

# Transition: Time

Static plot of trajectory for bat 1

```r
bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_point()
```
]
.right-plot[
<img src="slides_files/figure-html/timeout1-1.png" style="display: block; margin: auto;" />
]

---

# Transition: Time

Add `transition_time` with `time = time`

Duplicate points appearing in the same frame

```r
bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_point() +
* transition_time(time = time)
```
]
.right-plot[
<img src="slides_files/figure-html/timeout2-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: Time

Try adjusting the number of frames using `animate`

```r
*time_plot <- bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_point() +
 transition_time(time = time)

*animate(plot = time_plot,
* nframes = 300)
```
]
.right-plot[
<img src="slides_files/figure-html/timeout3-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: Time

Example of adjusting the time range

```r
bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_point() +
 transition_time(time = time, 
* range = c(1, 3))
```
]
.right-plot[
<img src="slides_files/figure-html/timeout4-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: Reveal

`transition_reveal`: Reveal data along a given *dimension*

Options:

- `along`: An unquoted expression giving the dimension to tween along. For a gradually revealing time series this should be set to the same as the x aesthetic.

- `range`: The time range to animate. If NULL it will be set to the range of along

- `keep_last = TRUE`: For non-path/polygon layers should the last row be kept for subsequent frames.

---

# Transition: Reveal

Static plot of trajectory for bat 1 (again)

```r
bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_point()
```
]
.right-plot[
<img src="slides_files/figure-html/revealout1-1.png" style="display: block; margin: auto;" />
]

---

# Transition: Reveal

Add `transition_reveal` with `along = time`

Note: No duplicated observations in a frame

```r
bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_point() +
* transition_reveal(along = time)
```
]
.right-plot[
<img src="slides_files/figure-html/revealout2-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: Reveal

`transition_time` vs `transition_reveal` (from the documentation):

> In contrast to e.g. transition_time() transition_reveal() calculates intermediary values at exact positions instead of  coercing raw values into the closest frame.

> It further keeps old data for path and polygon type layers so that they are gradually build up instead of being a set of disconnected segments...

---

# Transition: Reveal

Change to using `geom_path` instead of `geom_point` - allows data to gradually appear with `transition_reveal`

```r
bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
* geom_path() +
 transition_reveal(along = time) 
```
]
.right-plot[
<img src="slides_files/figure-html/revealout3-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: Reveal

If using both `geom_path` and `geom_point`, the history of the path will appear for the whole animation, but the point will only appear at the current time.

```r
bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
* geom_path() +
* geom_point() +
 transition_reveal(time)
```
]
.right-plot[
<img src="slides_files/figure-html/revealout4-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: Reveal

History of points can be kept by assigning the `along` variable to a group

```r
bats %>%
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_path() +
* geom_point(aes(group = time)) +
 transition_reveal(along = time)
```
]
.right-plot[
<img src="slides_files/figure-html/revealout5-1.gif" style="display: block; margin: auto;" />
]

---

# Transition: Reveal

Use the ggimage R package to replace the point with an image

```r
library(ggimage)

bat_image_link <- 
 "https://raw.githubusercontent.com/goodekat/presentations/master/2019-isugg-gganimate-spooky/figures/bat-cartoon.png"

bats %>%
 mutate(
* image = bat_image_link
 ) %>% 
 filter(id == 1) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_path() +
* geom_image(aes(image = image),
* size = 0.3) +
 transition_reveal(time)
```
]
.right-plot[
<img src="slides_files/figure-html/revealout6-1.gif" style="display: block; margin: auto;" />
]

---

# Your Turn

Use a transition to create the animation shown below.

If you want to use the same colors, add `scale_color_manual(values = bat_colors)` with `bat_colors <- c("darkorange", "orangered", "violetred", "purple", "black")`.

---

# Your Turn: Static Version

```r
bats %>%
  mutate(image = bat_image_link) %>%
  ggplot(aes(x = longitude, y = latitude, group = id, color = id)) +
  geom_path() +
  geom_image(aes(image = image), size = 0.1) +
  scale_color_manual(values = bat_colors)
```

---

# Your Turn: Solution

---

# Other Transitions

---

# Views

---

# View: Follow

Adjusts the axes to follow the data in the frame

```r
bats %>%
 mutate(image = bat_image_link) %>% 
 filter(id == 3) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_image(aes(image = image),
 size = 0.1) +
 geom_path() +
 transition_reveal(time) +
* view_follow()
```
]
.right-plot[
<img src="slides_files/figure-html/view_follow_out-1.gif" style="display: block; margin: auto;" />
]

---

# View: Step

Follows the data but switches between being static and changing to the range of the data

```r
bats %>%
 mutate(image = bat_image_link) %>% 
 filter(id == 3) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_image(aes(image = image),
 size = 0.1) +
 geom_path() +
 transition_reveal(time) +
* view_step(pause_length = 3,
* step_length = 1,
* nsteps = 5) #<<
```
]
.right-plot[
<img src="slides_files/figure-html/view_step_out1-1.gif" style="display: block; margin: auto;" />
]

---

# View: Step

Works will with `transistion_states`

```r
bats %>%
 ggplot(aes(x = longitude, 
 y = latitude, 
 color = time, 
 group = id)) +
 geom_point() + 
 scale_color_gradient2(
 midpoint = 6, 
 low = "orange", 
 mid = "purple",
 high = "black"
 ) + 
 transition_states(states = id) +
* view_step(pause_length = 3,
* step_length = 1,
* nsteps = 5)
```
]
.right-plot[
<img src="slides_files/figure-html/view_step_out2-1.gif" style="display: block; margin: auto;" />
]

---

# Other Views

| Function | Use |
| -------- | :----------------- |
| view_zoom | Pan and zoom smoothly between different states |
| view_static | Keep a fixed view that include all of the data |

---

# Your Turn

Apply a viewing option to the animation from the last 'your turn' to create an animation such as the one below.

---

# Your Turn: Solution

---

# Shadows

---

# Shadow: Wake

Leaves a "wake" behind the points as they move through time - options allow for the adjustment of length of wake, size of wake, and so forth

```r
bats %>%
 filter(id == 3) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_point() +
 transition_reveal(time) +
* shadow_wake(wake_length = 0.1,
* alpha = 0.5)
```
]
.right-plot[
<img src="slides_files/figure-html/shadow_wake_out-1.gif" style="display: block; margin: auto;" />
]

---

# Shadow: Trail

Leaves points behind as a memory of the trajectory the points followed (keeps every nth frame) - allows for the adjustments of the distance between points to remember, alpha of memory, and so forth

```r
bats %>%
 filter(id == 3) %>%
 ggplot(aes(x = longitude, 
 y = latitude)) +
 geom_point() +
 transition_reveal(time) +
* shadow_trail(distance = 0.01,
* alpha = 0.5,
* shape = 2)
```
]
.right-plot[
<img src="slides_files/figure-html/shadow_trail_out-1.gif" style="display: block; margin: auto;" />
]

---

# Other Shadows

| Function | Use |
| -------- | :----------------- |
| shadow_mark | Show original data as background marks |
| shadow_null | A non-existent shadow |

---

# Your Turn

Apply use a shadow option to create an animation such as the one below

---

# Your Turn: Solution

```r
bats %>%
  mutate(image = bat_image_link) %>%
  ggplot(aes(x = longitude, y = latitude, group = id, color = id)) +
  geom_point() +
  scale_color_manual(values = bat_colors) +
  transition_reveal(along = time) + 
  shadow_wake(wake_length = 0.2, size = 5, alpha = FALSE, colour = 'grey92')
```

---

# Other

---

# Enter/Exit

Specifies how the objects transition from one time to the next

.pull-left[
`enter_manual()`  
`enter_appear()`   
`enter_fade()`  
`enter_grow()`  
`enter_recolour()`  
`enter_recolor()`  
`enter_fly()`  
`enter_drift()`  
`enter_reset()` 
]

.pull-right[
`exit_manual()`  
`exit_disappear()`  
`exit_fade()`  
`exit_shrink()`   
`exit_recolour()`  
`exit_recolor()`   
`exit_fly()`   
`exit_drift()`  
`exit_reset()`  
]

---

# Enter/Exit

```r
bats %>%
 ggplot(aes(
 x = longitude, 
 y = latitude, 
 color = time, 
 group = id)) + 
 geom_point() + 
 scale_color_gradient2(
 midpoint = 6, 
 low = "orange", 
 mid = "purple",
 high = "black") +
 transition_states(
 states = id, 
 transition_length = 2, 
 state_length = 1) + 
* enter_fade() +
* exit_fade() +
* exit_fly(x_loc = -2.66300,
* y_loc = 51.40692)
```
]
.right-plot[
<img src="slides_files/figure-html/enter_exit_out-1.gif" style="display: block; margin: auto;" />
]

---

# ease_aes

"defines how the aesthetics associated with the ease during transitions"

- `quadratic`: Models a power-of-2 function
- `cubic`: Models a power-of-3 function
- `quartic`: Models a power-of-4 function
- `quintic`: Models a power-of-5 function
- `sine`: Models a sine function
- `circular`: Models a pi/2 circle arc
- `exponential`: Models an exponential function
- `elastic`: Models an elastic release of energy
- `back`: Models a pullback and release
- `bounce`: Models the bouncing of a ball

Modifiers: `-in`, `-out`, `-in-out`

Demonstrations of the easings: [https://easings.net/en](https://easings.net/en)

---

# ease_aes

Use `sine-in-out` to smooth the fade transitions

```r
bats %>%
 ggplot(aes(
 x = longitude, 
 y = latitude, 
 color = time, 
 group = id)) + 
 geom_point() + 
 scale_color_gradient2(
 midpoint = 6, 
 low = "orange", 
 mid = "purple",
 high = "black") +
 transition_states(
 states = id, 
 transition_length = 2, 
 state_length = 1) + 
 enter_fade() +
 exit_fade() +
* ease_aes('sine-in-out')
```
]
.right-plot[
<img src="slides_files/figure-html/ease_aes_out-1.gif" style="display: block; margin: auto;" />
]

---

# Rendering and Output

- `animate()`
- `print()` 
- `knit_print.gganim()`

Renderers provided by gganimate

- `gifski_renderer()` 
- `file_renderer()` 
- `av_renderer()` 
- `ffmpeg_renderer()` 
- `magick_renderer()` 
- `sprite_renderer()`
]

- `anim_save()`: Save an animation to a file
- `last_animation()`: Retrieve the last rendered animation
- `frame_vars()`: Access metadata about the frames in an animation
- `split_animation()`: Split an animation into chunks
]

---

# Resources

Official Website: https://gganimate.com/index.html

GitHub: https://github.com/thomasp85/gganimate

Talks on gganimate: https://gganimate.com/articles/extra/talks.html

Workshop by Mitchell O'Hara-Wild: https://github.com/numbats/gganimate-workshop