See aRtpack in Action • artpack

Introduction

So, it’s great to have different functions like rotator(), circle_data(), and art_pals(), but how exactly can these functions help beginners become generative artists? The inspiration behind artpack was to provide a way to bring the process of creating generative art data closer to R users who had always wanted to explore generative art but had never fully understood how to make it happen. This vignette will serve as an example of how to create a generative art piece with artpack that’s a bit more involved than just creating a single shape from scratch. Note that this isn’t meant to serve as a general tutorial on generative art, but rather a simple showcase of how one might use artpack to create and complete a generative art piece.

For Beginners

It can be discouraging to see established generative artists share their final works without revealing the coding process behind them. artpack addresses this by serving as both a tool for generative artists and a resource for learning. It provides the data needed to create specific geometric shapes while helping aspiring artists “think outside the ggplot grid” and enhance their generative art workflow.

For Seasoned Artists

artpack is meant to serve as a suite of convenience functions to be cherry-picked and used as needed.

The hope is that this vignette is helpful to all who are interested in artpack by seeing it “in action”.

Get Inspired

The first step in creating any art piece is to get inspired! For this vignette, we’ll draw inspiration from a classic retro sunset design featuring:

A large orange sun in the center.

Radiating sun rays across a gradient sky (made of blue and orange)

Stylized wavy layers below the sun representing water in teal and blue tones

An overlaid artistic pattern to give the sky a textured look

Remember, when searching for inspiration, the aim isn’t to replicate every image you come across pixel by pixel. Instead, it’s about interpreting and drawing inspiration from it to form a general idea of what we can create for our own image. In this instance, let’s craft a sunset with waves of water and a pattern of sun rays emerging from the sun. A helpful way to think about creating art with ggplot is to imagine that you’re building a collage from the “first” layer all the way from the back (background), to the “last” layer in the front (foreground).

For this image, let’s create five layers:

Layer 1 (Bottom): The gradient of colors that represent the sky.

Layer 2: An artistic overlay to give the sky some ‘depth’.

Layer 3: The sun rays that are coming from the sun.

Layer 4: The sun.

Layer 5 (Top): The waves of water.

Creating the Data

Layer 1 (The Sky)

To create a gradient, we’ll need a row of horizontal bars assigned to different colors. Using ggplot2’sgeom_segment() for this can save on memory (as opposed to creating the bars yourself). geom_segment() needs 4 values: x, xend, y, and yend. So we’ll create a data frame with those data points. We’ll use the art_pals() function to assign colors to the data we create. We also need to decide how many bars we want to have in the image, but before we do that, let’s determine the size of our canvas. To keep things simple, let’s make our x and y limits the same. c(0, 100).

# Library Load in-------
library(artpack)
library(tibble)
library(dplyr)
library(purrr)
library(ggplot2)

# Set canvas size#
x_lim <- c(0, 100)
y_lim <- x_lim

# Set seed for reproducibility
set.seed(1000)

Next, let’s see if we can get the perfect size bars to create the illusion of a sky gradient. Let’s try 100 to start. This means we’ll also need 100 colors to apply to each bar. art_pals() can be used for this. We’ll use the artpack color palette "beach" for the sky.

After we have the data frame together, we can start building our ggplot to see how the data is looking on the plot:

# Sky Specs----
n_sky_bars <- 100
vec_sky_colors <- art_pals("beach", n_sky_bars)
df_sky_bars <-
  tibble(
    x = x_lim[1], # Horizontal bars means the x value is constant
    xend = x_lim[2], # Horizontal bars means the xend values is constant
    y = seq(y_lim[1], y_lim[2], length = n_sky_bars),
    yend = y, # Horizontal bars means yend is the same as y
    color = vec_sky_colors
  )

# Start the ggplot
df_sky_bars |> 
  ggplot(aes(x,y)) +
  geom_segment(
    aes(xend = xend, yend = yend),
    color = df_sky_bars$color,
    linewidth = 4 #Tweak the bar thickness here if needed.
    ) +
  coord_equal(xlim = x_lim, ylim = y_lim, expand = FALSE)

Gradient sky background showing a smooth color transition from orange at the bottom to light blue at the top, created using horizontal bars with the artpack beach color palette.

Layer 2 (The Artistic Overlay)

This layer may look the most complex on the canvas, but can be done easily with artpack. For the overlay, let’s add some light, transparent swirls to give the sky some texture. We can use the packer() function for this:

# Make the data
df_overlay <-
  packer(
    n = 300,
    min_x = x_lim[1],
    max_x = x_lim[2],
    min_y = y_lim[1],
    max_y = y_lim[2],
    color_pal = "#F5E0CD",
    circle_type = "swirl"
  )
#> ⠙ Sampling for big-sized circles started
#> ✔ Sampling for big-sized circles started [1s]
#> 
#> ℹ Big-sized circles complete!
#> ✔ Big-sized circles complete! [15ms]
#> 
#> ⠙ Sampling for medium-sized circles started
#> ℹ Maximum sampling reached for medium-sized circles!
#> ⠙ Sampling for medium-sized circles started✔ Sampling for medium-sized circles started [4.4s]
#> 
#> ℹ Medium-sized circles complete!
#> ✔ Medium-sized circles complete! [13ms]
#> 
#> ⠙ Sampling for small-sized circles started
#> ✔ Sampling for small-sized circles started [1.2s]
#> 
#> ℹ Small-sized circles complete!
#> ✔ Small-sized circles complete! [35ms]

# And add it to the plot
df_sky_bars |>
  ggplot(aes(x,y)) +
  geom_segment(
    aes(xend = xend, yend = yend),
    color = df_sky_bars$color,
    linewidth = 4
  ) +
  geom_path(
    data = df_overlay,
    aes(group = group),
    color = "#ffffff",
    linewidth = .3,
    alpha = .09
  ) +
  coord_equal(xlim = x_lim, ylim = y_lim, expand = FALSE)

The same gradient sky background from orange to blue, now with a subtle white swirl pattern overlay added using artpack's packer function. The overlay creates thin, semi-transparent white curved lines that add texture and depth to the sky gradient.

Layer 3 (The Sun Rays)

For this layer, we’re going to make 1 triangle that will originate from the center of the sun. This will require a little bit of manual work and some trial and error, but after we have one triangle, we can easily iterate and rotate it to create the sun rays if we know how many rays we want:

# Sun rays Specs----
n_sun_rays <- 50
# The rays angles
vec_ray_angles <- seq(0, 180, length = n_sun_rays)
# The ray groups
vec_ray_groups <- group_numbers(vec_ray_angles, prefix = "ray", sep = "_")
# The fill colors
vec_ray_fills <- rep_along(vec_ray_groups, c("#F5E0CD", NA))
# Data for One sun ray
# using a tribble for manual shapes can be helpful to keep the correct coords together
df_sun_ray <-
  tribble(
    ~x, ~y,
    50, 10, # Where the center of the sun will be
    175, 10, #trial and error arbitrary x coord that covers the image fully
    175, 20,
    50, 10
  )

# Ray options to iterate through
lst_ray_opts <-
  list(
    vec_ray_angles,
    vec_ray_groups,
    vec_ray_fills
  )

# Create the data frame
# Note that purrr is this artist's preferred method of iteration, but you do you!
# You want those for-loops?? go ahead and get loopy!
df_sun_rays <-
  pmap(lst_ray_opts,
       ~df_sun_ray |>
         rotator(x, y, angle = ..1, anchor = c(50, 10)) |>
         mutate(
           group = ..2,
           fill = ..3
         )
  ) |>
  list_rbind()


# Add to the plot
df_sky_bars |>
  ggplot(aes(x,y)) +
  geom_segment(
    aes(xend = xend, yend = yend),
    color = df_sky_bars$color,
    linewidth = 4
  ) +
  geom_path(
    data = df_overlay,
    aes(group = group),
    color = "#ffffff",
    linewidth = .3,
    alpha = .09
  ) +
  geom_polygon(
    data = df_sun_rays,
    aes(group = group),
    fill = df_sun_rays$fill,
    alpha = .3
  ) +
  coord_equal(xlim = x_lim, ylim = y_lim, expand = FALSE)

The gradient sky with overlay, now featuring radiating sun rays. Multiple triangular light-colored rays emanate from a central point at coordinates (50, 10), creating a classic sunburst pattern across the upper portion of the image. The rays were created using artpack's rotator function to rotate and repeat a single triangle shape.

Layer 4.1 and 4.2* (The Sun)

This layer will be simple to make as it’s just a circle. Based on the coordinates for our sun rays, we know that the center of our sun will be at c(50, 10). We can also determine the size of the sun by setting the radius. We can also include some color options.

Because we want the outline of the sun to blend in with the sun rays from the previous step, we’ll actually add 2 layers of df_sun to the plot. One with the lightly transparent sun border that will blend in with the sun rays from the previous layer and with the filled in yellow color of the sun.

# Make a circle----
df_sun <-
  circle_data(
    x = 50,
    y = 10,
    r = 25,
    color = "#F5E0CD",
    fill = "#FAB649"
  )

# And add the two layers to the plot
df_sky_bars |>
  ggplot(aes(x,y)) +
  geom_segment(
    aes(xend = xend, yend = yend),
    color = df_sky_bars$color,
    linewidth = 4
  ) +
  geom_path(
    data = df_overlay,
    aes(group = group),
    color = "#ffffff",
    linewidth = .3,
    alpha = .09
  ) +
  geom_polygon(
    data = df_sun_rays,
    aes(group = group),
    fill = df_sun_rays$fill,
    alpha = .3
  ) +
  geom_path(
    data = df_sun,
    color = df_sun$color,
    alpha = .1,
    linewidth = 2
  ) +
  geom_polygon(
    data = df_sun,
    fill = df_sun$fill
  ) +
  coord_equal(xlim = x_lim, ylim = y_lim, expand = FALSE)

The layered artwork now includes a bright orange-yellow circular sun positioned at the center of the radiating sun rays. The sun was created using artpack's circle_data function and appears as a filled circle with a subtle light border that blends with the sun rays behind it.

Layer 5 (The Water Waves)

This last layer isn’t too bad, but does require some spec setup. Looking at the ggplot we have, we’d ideally want the “waves” of water to stay roughly below x = 10. We can pick the number of waves we want, then do some iteration with wave_data() to create an “ocean”:

# Wave specs----
# Number of waves
n_waves <- 400
# Starting points of the waves (randomized)
vec_wave_starts <- sample(seq(-75,75, l = 50), n_waves, replace = TRUE)
# Length of waves (randomized)
vec_wave_lengths <- vec_wave_starts + sample(seq(25,100, l = 30), n_waves, replace = TRUE)
# Sizes of waves (randomized)
vec_wave_sizes <- sample(seq(1, 3, l = 20), n_waves, replace = TRUE)
# frequency (humps) of each wave (randomized)
vec_wave_freqs <- sample(1:5, n_waves, replace = TRUE)
# Fill colors of waves
vec_wave_fills <- art_pals("ocean", n_waves, randomize = TRUE)
# Dampening of each wave
vec_wave_damps <- sample(seq(3,10, l = 10), n_waves, replace = TRUE)
# Group Numbering for each wave
vec_wave_groups <- group_numbers(1:n_waves)
# Wave Y transformations (randomized)
vec_wave_trans <- sample(seq(-5,9, l = 20), n_waves, replace = TRUE)

# Make the spec list
lst_wave_opts <-
  list(
    vec_wave_starts, # ..1
    vec_wave_lengths, #..2
    vec_wave_sizes, #..3
    vec_wave_freqs, #..4
    vec_wave_fills, #..5
    vec_wave_damps, #..6
    vec_wave_groups, #..7
    vec_wave_trans #..8
  )

# Make the data#
df_waves <-
  pmap(lst_wave_opts,
       ~wave_data(
         start = ..1,
         end = ..2, 
         size = ..3,
         freq = ..4,
         fill = ..5,
         dampen = ..6,
         group_var = TRUE
       ) |>
         mutate(
           group = paste0(..7),
           y = y + ..8)
  ) |>
  list_rbind()

# And add it to the plot
df_sky_bars |>
  ggplot(aes(x,y)) +
  geom_segment(
    aes(xend = xend, yend = yend),
    color = df_sky_bars$color,
    linewidth = 4
  ) +
  geom_path(
    data = df_overlay,
    aes(group = group),
    color = "#ffffff",
    linewidth = .3,
    alpha = .09
  ) +
  geom_polygon(
    data = df_sun_rays,
    aes(group = group),
    fill = df_sun_rays$fill,
    alpha = .3
  ) +
  geom_path(
    data = df_sun,
    color = df_sun$color,
    alpha = .1,
    linewidth = 2
  ) +
  geom_polygon(
    data = df_sun,
    fill = df_sun$fill
  ) +
  geom_polygon(
    data = df_waves,
    aes(group = group),
    fill = df_waves$fill,
    alpha = .5,
    position = position_jitter(width = .1, height = .02)
  ) +
  coord_equal(xlim = x_lim, ylim = y_lim, expand = FALSE) +
  theme_void()

# Use ggplot2 to save it locally if you'd like:
ggsave("vintage_sun.png", dpi = 300, bg = "transparent")

Wrapping Up

In this vignette, we demonstrated how artpack’s functions can work together to create complex generative art pieces. By thinking in layers and combining functions like art_pals(), circle_data(), wave_data(), and rotator(), you can build sophisticated art pieces step by step.

The key is starting with inspiration, breaking your vision into manageable layers, and iterating as you build. Each artpack function handles the mathematical and backend complexity of creating the art data, while letting you focus on the creative decisions.

Ready to create your own art pieces? Start with what inspired you and just jump in and have fun. For detailed documentation on each function, see help(package = "artpack"), visit the individual function help pages like ?circle_data(), ?wave_data(), and ?art_pals(), or visit the aRtpack reference page online here.