Partial Dependence Plot — Seaborn

A partial dependence plot (PDP) showing the marginal effect of one feature on the predicted outcome of a machine learning model. The plot displays how predictions change as a feature varies across its range, while averaging over the effects of all other features. This visualization is essential for understanding the relationship between individual features and model predictions in interpretable machine learning.

Partial Dependence Plot rendered with Seaborn

Python source (Seaborn)

""" anyplot.ai
pdp-basic: Partial Dependence Plot
Library: seaborn 0.13.2 | Python 3.13.13
Quality: 90/100 | Updated: 2026-05-15
"""

import os

import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns
from sklearn.datasets import make_regression
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.inspection import partial_dependence


# Theme tokens
THEME = os.getenv("ANYPLOT_THEME", "light")
PAGE_BG = "#FAF8F1" if THEME == "light" else "#1A1A17"
ELEVATED_BG = "#FFFDF6" if THEME == "light" else "#242420"
INK = "#1A1A17" if THEME == "light" else "#F0EFE8"
INK_SOFT = "#4A4A44" if THEME == "light" else "#B8B7B0"
BRAND = "#009E73"

# Set seed for reproducibility
np.random.seed(42)

# Generate synthetic regression data (housing price prediction scenario)
X, y = make_regression(n_samples=500, n_features=5, noise=10, random_state=42)

# Feature names for context
feature_names = ["Square Feet", "Bedrooms", "Age (years)", "Distance to City", "Lot Size"]

# Train a gradient boosting model
model = GradientBoostingRegressor(n_estimators=100, max_depth=4, random_state=42)
model.fit(X, y)

# Compute partial dependence for feature 0 (Square Feet)
feature_idx = 0
pd_result = partial_dependence(model, X, features=[feature_idx], kind="average", grid_resolution=80)

# Extract values
feature_values = pd_result["grid_values"][0]
pd_values = pd_result["average"][0]

# Center partial dependence at zero for easier interpretation
pd_values_centered = pd_values - pd_values.mean()

# Compute confidence interval using individual predictions
pd_individual = partial_dependence(model, X, features=[feature_idx], kind="individual", grid_resolution=80)
ice_lines = pd_individual["individual"][0]
ice_centered = ice_lines - ice_lines.mean(axis=1, keepdims=True)
std_dev = np.std(ice_centered, axis=0)
ci_lower = pd_values_centered - 1.96 * std_dev / np.sqrt(len(X))
ci_upper = pd_values_centered + 1.96 * std_dev / np.sqrt(len(X))

# Create figure with seaborn style
sns.set_theme(
    style="ticks",
    rc={
        "figure.facecolor": PAGE_BG,
        "axes.facecolor": PAGE_BG,
        "axes.edgecolor": INK_SOFT,
        "axes.labelcolor": INK,
        "text.color": INK,
        "xtick.color": INK_SOFT,
        "ytick.color": INK_SOFT,
        "grid.color": INK,
        "grid.alpha": 0.10,
    },
)

fig, ax = plt.subplots(figsize=(16, 9), facecolor=PAGE_BG)

# Plot confidence band
ax.fill_between(feature_values, ci_lower, ci_upper, alpha=0.25, color=BRAND, label="95% Confidence Interval")

# Plot main PDP line using seaborn
sns.lineplot(
    x=feature_values, y=pd_values_centered, ax=ax, color=BRAND, linewidth=3, label="Partial Dependence", legend=False
)

# Add rug plot to show data distribution
feature_data = X[:, feature_idx]
sns.rugplot(x=feature_data, ax=ax, color=INK_SOFT, height=0.03, alpha=0.5)

# Add horizontal line at zero for reference
ax.axhline(y=0, color=INK_SOFT, linestyle="--", linewidth=1.5, alpha=0.4)

# Styling
ax.set_xlabel(f"{feature_names[feature_idx]}", fontsize=20, color=INK)
ax.set_ylabel("Partial Dependence (centered)", fontsize=20, color=INK)
ax.set_title("pdp-basic · seaborn · anyplot.ai", fontsize=24, fontweight="medium", color=INK)
ax.tick_params(axis="both", labelsize=16, colors=INK_SOFT)

# Subtle grid
ax.yaxis.grid(True, alpha=0.2, linewidth=0.8, color=INK)
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)
for s in ("left", "bottom"):
    ax.spines[s].set_color(INK_SOFT)

# Legend
handles, labels = ax.get_legend_handles_labels()
ax.legend(
    handles,
    labels,
    fontsize=16,
    loc="upper left",
    frameon=True,
    fancybox=False,
    edgecolor=INK_SOFT,
    facecolor=ELEVATED_BG,
)

plt.tight_layout()
plt.savefig(f"plot-{THEME}.png", dpi=300, bbox_inches="tight", facecolor=PAGE_BG)

Part of Partial Dependence Plot on anyplot.ai.

Other implementations