Static plots are the bread and butter of R, but sometimes a still image doesn't capture the "story" of the data—especially when you're dealing with trends over time. That's where gganimate comes in. It's essentially a wrapper for ggplot2 that lets us treat time or categories as an animation frame.

For this lesson, I want us to build a visualization that tracks the hypothetical temperature trends of three different cities over a season. It's a simple scenario, but it'll let us touch on the most critical parts of the package: transitions, enter/exit animations, and the dreaded "jumpy line" bug.

Setting up our weather data

I'm going to generate a quick synthetic dataset. In a real project, you'd be loading a CSV here, but for this exercise, I want to make sure we have a clean time series to work with.

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

# Creating a fake dataset of 90 days for 3 cities
set.seed(42)
days <- 1:90
cities <- c("New York", "London", "Tokyo")

weather_data <- expand.grid(Day = days, City = cities) %>%
  group_by(City) %>%
  mutate(Temp = cumsum(rnorm(90, mean = 0.1, sd = 2)) + runif(1, 10, 20))

# Let's peek at it
head(weather_data)

Mapping the static foundation

Before we touch any animation functions, we have to build a standard ggplot. If the static plot doesn't look right, the animation will be a nightmare to debug. I'll start with a simple line chart. I'm using geom_line because we want to see the trajectory of the temperature.

# The base plot
p <- ggplot(weather_data, aes(x = Day, y = Temp, color = City)) +
  geom_line() +
  labs(title = "Temperature Trend: Day {frame_time}", 
       x = "Day of Season", 
       y = "Temperature (C)") +
  theme_minimal()

p

Adding the dimension of time

Now we turn this into an animation. The most common function you'll use is transition_time(). This tells R, "Instead of showing all the data at once, slice it by this variable and move through it sequentially."

I'll also add ease_aes(). Without it, the animation can feel robotic—linear movement is boring. ease_aes('cubic-in-out') makes the transition feel more organic, like it's accelerating and decelerating naturally.

anim <- p + 
  transition_time(Day) + 
  ease_aes('cubic-in-out')

# To view it, we call animate()
animate(anim, nframes = 100, fps = 10)

Correcting the "Jumpy Line" glitch

If you ran the code above, you might have noticed something weird: the lines aren't "drawing" themselves from left to right. Instead, they might be flickering or jumping around. I've made this mistake a dozen times.

The problem is that gganimate doesn't automatically know which points belong to which line when it's animating the transition. It treats every point as an individual entity. To fix this, we need to explicitly tell the aes() mapping that the lines are grouped by City.

# The Fix: Adding 'group = City' inside aes()
p_fixed <- ggplot(weather_data, aes(x = Day, y = Temp, color = City, group = City)) +
  geom_line() +
  labs(title = "Temperature Trend: Day {frame_time}", 
       x = "Day of Season", 
       y = "Temperature (C)") +
  theme_minimal()

anim_fixed <- p_fixed + 
  transition_time(Day) + 
  ease_aes('cubic-in-out')

animate(anim_fixed, nframes = 100, fps = 10)

Polishing the motion

To wrap this up, I want the lines to feel a bit more "alive." I'll add shadow_wake()` and exit_fade()`. The wake creates a trail behind the leading edge of the line, which helps the viewer track the movement. It's a small detail, but it's the difference between a "chart that moves" and a "professional visualization."

final_anim <- p_fixed + 
  transition_time(Day) + 
  shadow_wake(wake_length = 0.1, alpha = 0.4) +
  ease_aes('cubic-in-out')

animate(final_anim, nframes = 100, fps = 10)

📋 Practical Task

Exercise: Visualizing Regional Website Traffic Spikes

Your goal is to create an animated line chart showing website traffic for three different regions (North America, Europe, and Asia) over a 30-day period.

Requirements:

• Generate a data frame with columns: Day (1 to 30), Region (NA, EU, Asia), and Visitors (randomly generated numbers between 100 and 1000).
• Create a ggplot using geom_line. Ensure you use the group aesthetic to prevent the lines from jumping during animation.
• Use transition_time(Day) to animate the growth of the traffic.
• Include a dynamic title using {frame_time} to show which day is currently being displayed.
• Add a shadow_wake effect to the lines to make the movement smoother.
• Render the final animation with nframes = 60 and fps = 12.