Starburst Art

Using R to build beautiful digital artwork randomly

R is consistently lauded for its impressive data plotting capabilities. These capabilities can be co-opted to create beautiful art, programmatically. Below are a few examples of my results using a single core function that accepts many parameters and outputs a semi-randomly generated plot.

The general idea is to create:

One generic function that can plot segments with given inputs such as segment end positions and color values. Add a bit of randomness for effect. Shapes can be created by passing values to the segment positions. For example, sin(), cos(), and pi values can be used to create circle and semi circles
- this is buildFan()
A second function that wraps around the first to print multiple plots. It also allows parameters across the plots such as color gradients and shapes
- this is printFans()

If you’re more interested in this, I encourage you to tweak (and improve) my code along with checking out Data-to-Art.

# build a single fan
buildFan()

# tweak the fan's properties
buildFan(n.segments = 1000,
         x.end = runif(1000, -1, 1),
         y.end = runif(1000, -1, 1),
         x.center = rep(0, 1000),
         y.center = rep(0, 1000),
         alpha = 0.3,
         depth = TRUE)

You can create animated gifs by saving many plots and then collapse them into a GIF use imagemagick shell commands:

# gif rendering requires imagemagick to be installed
#  see https://imagemagick.org/ or brew install imagemagick 
setwd("Plotted_images/Gifs")

# gif of single fan
png(filename = paste0('fan%02d.png'),
    width = 500,
    height = 500)
for (i in c(1:40, 39:1)){
 
  # set the number of segments to plot
  n <- floor(i^2)

  buildFan(hue = i/40, 
           saturation = 1- i/40, 
           value = 0.5, 
           alpha = 1 - (i/40),
           n.segments = n,
           x.end = runif(n, -1, 1),
           y.end = runif(n, -1, 1),
           x.center = rep(0, n),
           y.center = rep(0, n),
           depth = TRUE)
}
dev.off()
system("convert -delay 20 *.png single_fan.gif")
file.remove(list.files(pattern = ".png"))

# gif parameters
n.frames <- 100
logis.curve <- 1 / ( 1 + exp(seq(4, -5, length = n.frames/2)))
logis.curve <- c(logis.curve, rev(logis.curve))
rescaleVector <- function(x, min = .1, max = .9) (x - min(x))/(max(x) - min(x)) * (max - min) + min

# set where to save
png(filename = paste0('fan%02d.png'),
    width = 1250,
    height = 1250)
for (i in 1:n.frames){
  
  # set the number of segments to plot
  n <- floor(logis.curve[i]*250)
  
  buildFan(hue = rescaleVector(logis.curve)[i], 
           saturation = .6, 
           value = 0.5, 
           hue.range = 0.20,
           saturation.range = 0.20,
           value.range = 0.20,
           alpha =  1 - rescaleVector(logis.curve)[i],
           alpha.range = 0.10,
           segment.width = 6,
           n.segments = n,
           x.end = runif(n, -1, 1),
           y.end = runif(n, -1, 1),
           x.center = rep(0, n),
           y.center = rep(0, n),
           depth = TRUE
  )
}
dev.off()
system("convert -delay 20 *.png depth_fan.gif")
file.remove(list.files(pattern = ".png"))

# set where to save
png(filename = paste0('fan%02d.png'),
    width = 1250,
    height = 800)

for (i in 1:n.frames){
  
  # set the number of segments to plot
  n <- floor(logis.curve[i]*400)
  
  # semi circle
  base.vector <- seq(0, pi, length = n)
  x.vector <- cos(base.vector)
  y.vector <- sin(base.vector) * 2 - 1
  
  buildFan(hue = rescaleVector(logis.curve, min = 0.3, max = .6)[i], 
           saturation = .45, 
           value = 0.4, 
           hue.range = 0.20,
           saturation.range = 0.20,
           value.range = 0.20,
           alpha =  1 - rescaleVector(logis.curve, min = .6, max = .95)[i],
           alpha.range = 0.30,
           segment.width = 20,
           n.segments = n,
           x.end = x.vector,
           y.end = y.vector,
           x.center = 0,
           y.center = -1
  )
}
dev.off()
system("convert -delay 20 *.png semi_fan.gif")
file.remove(list.files(pattern = ".png"))

You can create animated gifs across plots, specifying individual parameters to create effects.

# gif of many fans with sequencing pattern
# gif parameters
n.frames <- 12*8
hues <- rep(seq(0, 1, length = 12), 12)
logis.curve <- 1 / ( 1 + exp(seq(4, -4, length = n.frames/2)))
logis.curve <- c(logis.curve, rev(logis.curve))
loop.vec <- c(10:11, 0:11, 0)

# set where to save
png(filename = 'fan%02d.png',
    width = 1250,
    height = 1250)

for (i in 1:n.frames){

  # set graphical parameters
  par(bg = 'white',
      mar = rep(2, 4),
      mfrow = c(3, 4),
      bty = "n")
  
  # set the number of segments to plot
  n <-  floor(logis.curve[i] * 250)
  
  for (j in 1:12) {
    
    # logic controls the plot sequence
    if (i %% 12 %in% loop.vec[(j+3)-0:3]){
      buildFan(
        hue = hues[j + i],
        saturation = .7,
        value = 0.5,
        alpha = 1 - logis.curve[i],
        n.segments = n,
        x.end = runif(n, -1, 1),
        y.end = runif(n, -1, 1),
        # x.center = rep(0, n),
        # y.center = rep(0, n),
        depth = TRUE,
        random.center = TRUE
      )
    } else {
      # build a blank plot
      buildFan(alpha = 0)
    }
  }
}
dev.off()
system("convert -delay 40 *.png multi_fan.gif")
file.remove(list.files(pattern = ".png"))

Use printFans() to create a panel of plots all at once.

# build a panel of fans and save it
printFans()

Tweaking the parameters leads to many different effects