Random Walk of Pi – Another ggplot2 Experiment

There are so many beautiful “π” arts everywhere, and I wanted to practice ggplot2 by mimicing those arts further more. Another pi art caught my eye is random walk of pi digits. Here’s one of examples in WIRED magazine.

For random walk to work, I’ve assigned direction to “walk” depending on digits 0-9.

I’ve prepared data frame as below, so I can use ggplot2 to plot

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
## pi_df contains first 100K digits , each digit is stored in one row.

df_walk <- pi_df %>% 
 mutate(pos = row_number()-1, ## assign position
 dig= as.numeric(dig)) %>%
 select(pos, dig) %>%
 mutate(angle_rad = 2*pi/10*dig, ## using current digit determine direction to move
 angle_deg = circular::deg(angle_rad), ## I just like to see number in degree...
 move_x = cos(angle_rad), ## how much to move in x direction
 move_y = sin(angle_rad), ## how much to move in y direction
 last_x = replace_na(lag(move_x),0), ## position of last x, set origin as 0
 last_y = replace_na(lag(move_y),0), ## position of last y, set origin as 0
 cumsum_x = cumsum(move_x), ## walkig == adding up all steps in x
 cumsum_y = cumsum(move_y), ## walking == adding up all steps in y
 cumsum_x_lag = cumsum(last_x),
 cumsum_y_lag = cumsum(last_y)) 

df_walk %>% head(n=5) %>% knitr::kable()
0 3 1.8849556 108 -0.309017 0.9510565 0.000000 0.0000000 -0.309017 0.9510565 0.000000 0.0000000
1 1 0.6283185 36 0.809017 0.5877853 -0.309017 0.9510565 0.500000 1.5388418 -0.309017 0.9510565
2 4 2.5132741 144 -0.809017 0.5877853 0.809017 0.5877853 -0.309017 2.1266270 0.500000 1.5388418
3 1 0.6283185 36 0.809017 0.5877853 -0.809017 0.5877853 0.500000 2.7144123 -0.309017 2.1266270
4 5 3.1415927 180 -1.000000 0.0000000 0.809017 0.5877853 -0.500000 2.7144123 0.500000 2.7144123

Now to visualize the random walk, I’ve used below script to visualize first 1000, first 10000 and first 100000 digits of pi.

1
2
3
4
5
6
7
8
9
10
11
12
13
n_steps <- 1000
## Random Walk of Pi
df_walk %>% 
 filter(pos < n_steps) %>%
 ggplot(aes(x=cumsum_x, y=cumsum_y, color=pos)) + 
 geom_segment(size=0.5, aes(xend=cumsum_x_lag, yend=cumsum_y_lag)) +
 geom_point(size=0.8) + 
 theme_void(base_family="Roboto Condensed") +
 theme(panel.background=element_rect(fill="#000000")) +
 scale_color_viridis_c(option="plasma", guide="none") +
 labs(caption=paste("First",n_steps,"digits of Pi")) +
 geom_hline(yintercept=0, color="#ffffff30", linetype=3) +
 geom_vline(xintercept=0, color="#ffffff30", linetype=3)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
#ggsave(str_c("output/random_walk_pi_",n_steps,".png"), width=11, height=7)


n_steps <- 10000
## Random Walk of Pi
df_walk %>% 
 filter(pos < n_steps) %>%
 ggplot(aes(x=cumsum_x, y=cumsum_y, color=pos)) + 
 geom_segment(size=0.1, aes(xend=cumsum_x_lag, yend=cumsum_y_lag)) +
 geom_point(size=0.01) + 
 theme_void(base_family="Roboto Condensed") +
 theme(panel.background=element_rect(fill="#000000")) +
 scale_color_viridis_c(option="plasma", guide="none") +
 labs(caption=paste("First",n_steps,"digits of Pi")) +
 geom_hline(yintercept=0, color="#ffffff30", linetype=3) +
 geom_vline(xintercept=0, color="#ffffff30", linetype=3)

1
2
3
4
5
6
7
8
9
10
11
12
13
n_steps <- 100000
## Random Walk of Pi
df_walk %>% 
 filter(pos < n_steps) %>%
 ggplot(aes(x=cumsum_x, y=cumsum_y, color=pos)) + 
 geom_segment(size=0.1, aes(xend=cumsum_x_lag, yend=cumsum_y_lag)) +
 geom_point(size=0.005) + 
 theme_void(base_family="Roboto Condensed") +
 theme(panel.background=element_rect(fill="#000000")) +
 scale_color_viridis_c(option="plasma", guide="none") +
 labs(caption=paste("First 100000 digits of Pi")) +
 geom_hline(yintercept=0, color="#ffffff30", linetype=3) +
 geom_vline(xintercept=0, color="#ffffff30", linetype=3)

Related