wiki: Random walk animation

[kevinsoo]

Using ggplot2 and gganimate to visualize simple random walks, used to determine draft order in my Fantasy Football league.

Create data

Random walks for 4 entities for a predetermined number of steps. Each step, I draw from a normal distribution with mean of 1, so that numbers are generally positive. The cumulative sum will be plotted, to show which entity gains the highest score during the walk.

# Load libraries
library(tidyverse)
library(gganimate)

# Generate data
n <- 20 # Number of iterations/steps

# Random walk for 4 players
df <- tibble(
    ite = 0:n,
    Player1 = c(0, rnorm(n, mean = 1)),
    Player2 = c(0, rnorm(n, mean = 1)),
    Player3 = c(0, rnorm(n, mean = 1)),
    Player4 = c(0, rnorm(n, mean = 1))) %>%
    gather(player, score, Player1:Player4) %>%
    arrange(player, ite) %>%
    group_by(player) %>%
    mutate(totalScore = cumsum(score),
           size = ifelse(ite == n, 6, 4)) # This is used to size the labels

Plot data

Plot uses labels to represent each entity, and becomes larger at the final step so the final order is clearly visible.

# Plot data
ggplot(df, aes(x = player, y = totalScore)) +
    geom_hline(yintercept = 0, linetype ="dashed") +
    geom_label(aes(label = player, fill = player, size = size)) +
    theme_minimal() +
    theme(legend.position = 'none') +
    scale_fill_brewer(palette = "Spectral") +
    labs(title = "Random walk race",
         subtitle = "Iteration: {closest_state} of 20",
         y = "Score total",
         x = "Entity",
         caption = "Random walk simulated for 20 iterations") +
    transition_states(ite, transition_length = 1, state_length = 1, wrap = FALSE) +
    ease_aes('linear') +
    shadow_wake(wake_length = .1, wrap = FALSE)

[thomasp85]

https://github.com/thomasp85/gganimate/wiki/Football-Draft-Order