Animate ggplot Time Series Plot with a Sliding Window using R

Animating time series data with a sliding window in R is a visually engaging way to represent changes over time, especially in trends or forecasts. The gganimate package builds on the ggplot2 framework to create smooth, animated visualizations, and by using a sliding window approach, we can focus on a rolling subset of data. In this article, we will walk through the steps of animating a time series plot with a sliding window in R Programming Language.

Make sure to install the necessary packages:

library(ggplot2)
library(gganimate)
library(dplyr)

Step 1: Simulating a Time Series Dataset

To demonstrate the sliding window animation, we will generate a time series dataset using the seq() function for time points and rnorm() for random normal values representing observations.

# Create a simulated time series dataset
set.seed(123)
n <- 100
time <- seq(as.Date("2022-01-01"), by = "day", length.out = n)
value <- cumsum(rnorm(n, mean = 0.2, sd = 1))  # Random walk

time_series_data <- data.frame(time, value)
head(time_series_data)

Output:

        time      value
1 2022-01-01 -0.3604756
2 2022-01-02 -0.3906531
3 2022-01-03  1.3680552
4 2022-01-04  1.6385636
5 2022-01-05  1.9678513
6 2022-01-06  3.8829163

• time: Date values representing time points.
• value: Randomly generated values representing the observations at each time point.

Step 2: Visualizing the Time Series Plot

First, let’s create a simple static ggplot2 line plot to visualize the entire time series.

# Basic time series plot
ggplot(time_series_data, aes(x = time, y = value)) +
  geom_line(color = "blue") +
  labs(title = "Time Series Plot",
       x = "Time",
       y = "Value") +
  theme_minimal()

Output:

This static plot gives us an overview of the time series data. Now, let’s animate it with a sliding window effect.

Step 3: Adding a Sliding Window

We will use the gganimate package to animate the plot by showing only a sliding window of data that changes over time. Let’s define a window size of 20 observations. The sliding window will only show these observations at any given frame in the animation.

# Define the window size
window_size <- 20

# Create a new column to indicate frame number for the animation
time_series_data <- time_series_data %>%
  mutate(frame = row_number())  # Frame number corresponds to row number

Step 4: Creating the Animated Plot

We will create an animated plot where only a subset of data within the sliding window is displayed at each frame.

# Create the animated time series plot with a sliding window
animated_plot <- ggplot(time_series_data, aes(x = time, y = value)) +
  geom_line(color = "blue") +
  labs(title = 'Time Series Animation: {frame_time}',
       x = "Time",
       y = "Value") +
  theme_minimal() +
  
  # Animate the sliding window by showing only the last 'window_size' points
  transition_time(frame) +
  view_follow(fixed_y = TRUE) +
  
  # Limit the number of points shown in each frame
  enter_fade() +
  exit_fade() +
  
  shadow_mark(past = TRUE, future = FALSE, alpha = 0.2)

# Render the animation
animate(animated_plot, nframes = n - window_size, fps = 10)

Output:

Rendering [============================================>-------] at 8.7 fps ~ eta:  1s`
ℹ Do you need to adjust the group aesthetic?
Rendering [=============================================>------] at 8.7 fps ~ eta:  1s`
ℹ Do you need to adjust the group aesthetic?
`geom_line()`: Each group consists of only one observation.
ℹ Do you need to adjust the group aesthetic?
Rendering [==============================================>-----] at 8.7 fps ~ eta:  1s`
ℹ Do you need to adjust the group aesthetic?
`geom_line()`: Each group consists of only one observation.
ℹ Do you need to adjust the group aesthetic?
Rendering [===============================================>----] at 8.7 fps ~ eta:  
ℹ Do you need to adjust the group aesthetic?
Rendering [================================================>---] at 8.7 fps ~ eta:  
ℹ Do you need to adjust the group aesthetic?
Rendering [================================================>---] at 8.7 fps ~ eta:  
ℹ Do you need to adjust the group aesthetic?
Rendering [=================================================>--] at 8.7 fps ~ eta:  
ℹ Do you need to adjust the group aesthetic?
Rendering [==================================================>-] at 8.7 fps ~ eta:  0s`
ℹ Do you need to adjust the group aesthetic?
`geom_line()`: Each group consists of only one observation.
ℹ Do you need to adjust the group aesthetic?
# A tibble: 80 × 7                                                                    
   format width height colorspace matte filesize density
   <chr>  <int>  <int> <chr>      <lgl>    <int> <chr>  
 1 gif      480    480 sRGB       FALSE        0 72x72  
 2 gif      480    480 sRGB       TRUE         0 72x72  
 3 gif      480    480 sRGB       TRUE         0 72x72  
 4 gif      480    480 sRGB       TRUE         0 72x72  
 5 gif      480    480 sRGB       TRUE         0 72x72  
 6 gif      480    480 sRGB       TRUE         0 72x72  
 7 gif      480    480 sRGB       TRUE         0 72x72  
 8 gif      480    480 sRGB       TRUE         0 72x72  
 9 gif      480    480 sRGB       TRUE         0 72x72  
10 gif      480    480 sRGB       TRUE         0 72x72 

• transition_time(): This function is used to transition between frames based on the frame column, which represents each point in time.
• view_follow(fixed_y = TRUE): Ensures that the y-axis limits follow the values within the window. fixed_y = TRUE ensures that the y-axis limits stay consistent across frames.
• shadow_mark(): This function highlights the trailing points within the sliding window, giving a smooth fading effect for points outside the window.
• enter_fade() and exit_fade(): Controls how points appear and disappear in the animation.

The resulting animation will show a sliding window of 20 points moving along the time axis.

Conclusion

Animating a time series plot with a sliding window using gganimate provides an insightful way to visualize trends in your data over time. In this guide, we covered how to:

• Create a static time series plot using ggplot2.
• Animate the plot with a sliding window using gganimate.