Ternary Density Plot — Seaborn

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.
Ternary Density Plot rendered with Seaborn
Python source (Seaborn)

""" anyplot.ai
ternary-density: Ternary Density Plot
Library: seaborn 0.13.2 | Python 3.13.13
Quality: 91/100 | Updated: 2026-05-19
"""

import os

import matplotlib.cm as cm
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns
from matplotlib.colors import Normalize
from matplotlib.patches import Polygon


# Theme tokens
THEME = os.getenv("ANYPLOT_THEME", "light")
PAGE_BG = "#FAF8F1" if THEME == "light" else "#1A1A17"
INK = "#1A1A17" if THEME == "light" else "#F0EFE8"
INK_SOFT = "#4A4A44" if THEME == "light" else "#B8B7B0"
ACCENT = "#4467A3"  # Okabe-Ito blue — structural elements

sns.set_theme(style="ticks", rc={"figure.facecolor": PAGE_BG, "axes.facecolor": PAGE_BG, "text.color": INK})

# Data - Soil composition samples (sand/silt/clay percentages)
np.random.seed(42)

# Cluster 1: Sandy soils (high sand content)
n1 = 200
sand1 = np.random.beta(5, 2, n1) * 70 + 25
silt1 = np.random.beta(2, 3, n1) * (100 - sand1) * 0.6
clay1 = 100 - sand1 - silt1

# Cluster 2: Silty soils (high silt content)
n2 = 150
silt2 = np.random.beta(5, 2, n2) * 60 + 30
sand2 = np.random.beta(2, 3, n2) * (100 - silt2) * 0.5
clay2 = 100 - sand2 - silt2

# Cluster 3: Clay-rich soils
n3 = 150
clay3 = np.random.beta(4, 2, n3) * 50 + 30
sand3 = np.random.beta(2, 3, n3) * (100 - clay3) * 0.4
silt3 = 100 - clay3 - sand3

sand = np.concatenate([sand1, sand2, sand3])
silt = np.concatenate([silt1, silt2, silt3])
clay = np.concatenate([clay1, clay2, clay3])

# Transform ternary to Cartesian: Sand → bottom-left, Silt → bottom-right, Clay → top
total = sand + silt + clay
sand_norm = sand / total
silt_norm = silt / total
clay_norm = clay / total
x = 0.5 * (2 * silt_norm + clay_norm)
y = (np.sqrt(3) / 2) * clay_norm

sqrt3_2 = np.sqrt(3) / 2
vertices = np.array([[0, 0], [1, 0], [0.5, sqrt3_2]])

# Plot
fig, ax = plt.subplots(figsize=(12, 12), facecolor=PAGE_BG)
ax.set_facecolor(PAGE_BG)

triangle_clip = Polygon(vertices, transform=ax.transData)

# Grid lines at 10% intervals (all three ternary families)
for i in range(1, 10):
    frac = i / 10
    gkw = {"color": INK_SOFT, "alpha": 0.20, "linewidth": 0.8, "zorder": 1}
    # Constant clay (horizontal, parallel to base)
    ax.plot([0.5 * frac, 1 - 0.5 * frac], [sqrt3_2 * frac, sqrt3_2 * frac], **gkw)
    # Constant silt (parallel to Sand-Clay left edge, slope +√3)
    ax.plot([frac, 0.5 * (1 + frac)], [0, sqrt3_2 * (1 - frac)], **gkw)
    # Constant sand (parallel to Silt-Clay right edge, slope -√3)
    ax.plot([1 - frac, 0.5 * (1 - frac)], [0, sqrt3_2 * (1 - frac)], **gkw)

# KDE density fill
sns.kdeplot(x=x, y=y, fill=True, cmap="viridis", levels=20, alpha=0.85, ax=ax, thresh=0.02, zorder=5)

for collection in ax.collections:
    collection.set_clip_path(triangle_clip)

# KDE contour lines — thicker for visibility at full resolution
sns.kdeplot(x=x, y=y, levels=10, color=ACCENT, linewidths=3.0, ax=ax, zorder=6)

for collection in ax.collections:
    collection.set_clip_path(triangle_clip)

# Triangle boundary
ax.add_patch(Polygon(vertices, fill=False, edgecolor=ACCENT, linewidth=4, zorder=15))

# Colorbar showing relative density scale
sm = cm.ScalarMappable(cmap="viridis", norm=Normalize(vmin=0, vmax=1))
sm.set_array([])
cbar = plt.colorbar(sm, ax=ax, fraction=0.025, pad=0.04, aspect=20, shrink=0.55)
cbar.set_label("Relative Density", fontsize=18, labelpad=12)
cbar.ax.yaxis.label.set_color(INK)
cbar.set_ticks([0, 0.5, 1.0])
cbar.ax.set_yticklabels(["Low", "Medium", "High"])
for lbl in cbar.ax.get_yticklabels():
    lbl.set_color(INK_SOFT)
    lbl.set_fontsize(16)
cbar.ax.tick_params(colors=INK_SOFT)
cbar.outline.set_edgecolor(INK_SOFT)

# Vertex labels (component names + unit)
ax.text(0, -0.08, "Sand (%)", ha="center", va="top", fontsize=22, fontweight="bold", color=INK)
ax.text(1, -0.08, "Silt (%)", ha="center", va="top", fontsize=22, fontweight="bold", color=INK)
ax.text(0.5, sqrt3_2 + 0.08, "Clay (%)", ha="center", va="bottom", fontsize=22, fontweight="bold", color=INK)

# Percentage tick labels along each edge
for i in [2, 4, 6, 8]:
    frac = i / 10
    ax.text(frac, -0.04, f"{int(frac * 100)}", ha="center", va="top", fontsize=16, color=INK_SOFT)
    ax.text(
        0.5 * frac - 0.04, sqrt3_2 * frac, f"{int(frac * 100)}", ha="right", va="center", fontsize=16, color=INK_SOFT
    )
    ax.text(
        1 - 0.5 * frac + 0.04, sqrt3_2 * frac, f"{int(frac * 100)}", ha="left", va="center", fontsize=16, color=INK_SOFT
    )

# Cluster annotations — label each density peak for data storytelling
ann_kw = {
    "fontsize": 15,
    "fontweight": "bold",
    "color": INK,
    "bbox": {"boxstyle": "round,pad=0.3", "facecolor": PAGE_BG, "edgecolor": INK_SOFT, "alpha": 0.75},
}
ax.text(0.13, 0.04, "Sandy\nsoils", ha="center", va="center", **ann_kw)
ax.text(0.80, 0.11, "Silty\nsoils", ha="center", va="center", **ann_kw)
ax.text(0.53, 0.50, "Clay-rich\nsoils", ha="center", va="center", **ann_kw)

# Style
ax.set_title(
    "Soil Composition · ternary-density · python · seaborn · anyplot.ai",
    fontsize=24,
    fontweight="medium",
    color=INK,
    pad=25,
)
ax.set_xlim(-0.15, 1.15)
ax.set_ylim(-0.15, 1.05)
ax.set_aspect("equal")
ax.axis("off")

plt.tight_layout()
plt.savefig(f"plot-{THEME}.png", dpi=300, bbox_inches="tight", facecolor=PAGE_BG)
Part of Ternary Density Plot on anyplot.ai.
Ternary Density Plot — Seaborn

Python source (Seaborn)

Other implementations
