A ternary density plot combines a three-component ternary diagram with kernel density estimation to visualize where compositional data concentrates. Instead of showing individual points, this visualization uses a heatmap overlay to reveal the underlying probability distribution of compositions, making it ideal for identifying clusters, modes, and patterns in large compositional datasets.
# anyplot.ai
# ternary-density: Ternary Density Plot
# Library: makie 0.21.9 | Julia 1.11.9
# Quality: 91/100 | Created: 2026-09-02
using CairoMakie
using Colors
using Random
using Statistics
Random.seed!(42)
# --- Theme tokens -----------------------------------------------------------
const THEME = get(ENV, "ANYPLOT_THEME", "light")
const PAGE_BG = THEME == "light" ? colorant"#FAF8F1" : colorant"#1A1A17"
const INK = THEME == "light" ? colorant"#1A1A17" : colorant"#F0EFE8"
const INK_SOFT = THEME == "light" ? colorant"#4A4A44" : colorant"#B8B7B0"
const GRID_RGBA = RGBAf(INK.r, INK.g, INK.b, 0.35)
# Imprint sequential colormap — single-polarity continuous data (density)
const IMPRINT_SEQ = cgrad([colorant"#009E73", colorant"#4467A3"])
# --- Data: simulated sediment composition (sand / silt / clay), two facies --
n_samples = 2400
n_sandy = round(Int, 0.55 * n_samples)
component_a = zeros(n_samples) # sand
component_b = zeros(n_samples) # silt
component_c = zeros(n_samples) # clay
for i in 1:n_samples
shape_a, shape_b, shape_c = i <= n_sandy ? (7, 3, 2) : (2, 3, 7)
gamma_a = -sum(log.(rand(shape_a)))
gamma_b = -sum(log.(rand(shape_b)))
gamma_c = -sum(log.(rand(shape_c)))
total = gamma_a + gamma_b + gamma_c
component_a[i] = gamma_a / total
component_b[i] = gamma_b / total
component_c[i] = gamma_c / total
end
# Barycentric -> Cartesian (A at (0,0), B at (1,0), C at (0.5, sqrt(3)/2))
tri_height = sqrt(3) / 2
x_coords = component_b .+ component_c .* 0.5
y_coords = component_c .* tri_height
# --- Kernel density estimate on a fine grid, masked to the triangle --------
bandwidth = mean([std(x_coords), std(y_coords)]) * n_samples^(-1 / 6)
# Bin edges exactly at the triangle's bounding box, so the heatmap below has no
# half-cell overhang past [0, 1] x [0, tri_height] (Makie centers cells otherwise).
edge_xs = range(0, 1; length = 161)
edge_ys = range(0, tri_height; length = 141)
center_xs = (edge_xs[1:end-1] .+ edge_xs[2:end]) ./ 2
center_ys = (edge_ys[1:end-1] .+ edge_ys[2:end]) ./ 2
density_grid = Array{Float64}(undef, length(center_xs), length(center_ys))
for (i, gx) in enumerate(center_xs), (j, gy) in enumerate(center_ys)
weights = exp.(-((gx .- x_coords) .^ 2 .+ (gy .- y_coords) .^ 2) ./ (2 * bandwidth^2))
density_grid[i, j] = sum(weights) / (n_samples * 2 * pi * bandwidth^2)
end
contour_levels = maximum(density_grid) .* [0.15, 0.35, 0.55, 0.75]
# --- Figure -------------------------------------------------------------------
title_str = "Sediment Composition · ternary-density · julia · makie · anyplot.ai"
fig = Figure(
size = (1600, 900),
fontsize = 14,
backgroundcolor = PAGE_BG,
)
ax = Axis(
fig[1, 1];
title = title_str,
titlesize = 20,
titlecolor = INK,
backgroundcolor = PAGE_BG,
aspect = DataAspect(),
)
hidedecorations!(ax)
hidespines!(ax)
# Ternary grid lines (beneath the density layer), every 20% of each component
for k in (0.2, 0.4, 0.6, 0.8)
# constant component_a = k, parallel to edge B-C
lines!(ax, [1 - k, (1 - k) * 0.5], [0, (1 - k) * tri_height]; color = GRID_RGBA, linewidth = 1.2)
# constant component_b = k, parallel to edge A-C
lines!(ax, [k, k + (1 - k) * 0.5], [0, (1 - k) * tri_height]; color = GRID_RGBA, linewidth = 1.2)
# constant component_c = k, parallel to edge A-B
lines!(ax, [k * 0.5, 1 - k * 0.5], [k * tri_height, k * tri_height]; color = GRID_RGBA, linewidth = 1.2)
end
# Density heatmap over the grid's bounding box (Imprint sequential)
hm = heatmap!(ax, edge_xs, edge_ys, density_grid; colormap = IMPRINT_SEQ, alpha = 0.92)
# Contour lines at key density levels for easier interpretation
contour!(ax, center_xs, center_ys, density_grid; levels = contour_levels, color = INK, linewidth = 1.3, alpha = 0.55)
# Mask the two bounding-box corners outside the triangle with crisp vector edges
# (avoids the staircase artifact a rectilinear heatmap grid would leave on the diagonals)
poly!(ax, Point2f[(0, 0), (0, tri_height), (0.5, tri_height)]; color = PAGE_BG, strokewidth = 0)
poly!(ax, Point2f[(1, 0), (1, tri_height), (0.5, tri_height)]; color = PAGE_BG, strokewidth = 0)
# Triangle outline
lines!(ax, [0, 1, 0.5, 0], [0, 0, tri_height, 0]; color = INK_SOFT, linewidth = 2.5)
# Vertex labels
text!(ax, 0, -0.05; text = "Sand", align = (:center, :top), color = INK, fontsize = 16)
text!(ax, 1, -0.05; text = "Silt", align = (:center, :top), color = INK, fontsize = 16)
text!(ax, 0.5, tri_height + 0.05; text = "Clay", align = (:center, :bottom), color = INK, fontsize = 16)
# Explicit limits — automatic axis limits don't account for text bounding boxes,
# which would otherwise clip the vertex labels against the scene edge.
xlims!(ax, -0.1, 1.1)
ylims!(ax, -0.13, tri_height + 0.13)
Colorbar(fig[1, 2], hm; label = "Density", labelcolor = INK, ticklabelcolor = INK_SOFT, labelsize = 14, ticklabelsize = 12)
colsize!(fig.layout, 1, Relative(0.88))
# --- Save -------------------------------------------------------------------
save("plot-$(THEME).png", fig; px_per_unit = 2)
Runnable source as JSON, for any HTTP client: https://api.anyplot.ai/specs/ternary-density/makie/code. Any spec id and library id listed in llms-full.txt fit the same URL shape; every URL below is complete and callable.
{
"spec_id": "ternary-density",
"language": "julia",
"library": "makie",
"page": "https://anyplot.ai/ternary-density/julia/makie",
"hub": "https://anyplot.ai/ternary-density",
"code_json": "https://api.anyplot.ai/specs/ternary-density/makie/code",
"spec_json": "https://api.anyplot.ai/specs/ternary-density",
"render_light_png": "https://storage.googleapis.com/anyplot-images/plots/ternary-density/julia/makie/plot-light.png",
"render_dark_png": "https://storage.googleapis.com/anyplot-images/plots/ternary-density/julia/makie/plot-dark.png",
"quality_score": 91.0,
"license": "MIT",
"guide": "https://anyplot.ai/llms.txt"
}Part of Ternary Density Plot on anyplot.ai.