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.
""" anyplot.ai
pdp-basic: Partial Dependence Plot
Library: bokeh 3.9.0 | Python 3.13.13
Quality: 90/100 | Updated: 2026-05-15
"""
import importlib
import sys
from pathlib import Path
# Remove current directory from sys.path to avoid shadowing bokeh package
sys.path = [p for p in sys.path if Path(p).resolve() != Path(__file__).resolve().parent]
# Import bokeh module and its submodules
bokeh_io = importlib.import_module("bokeh.io")
bokeh_models = importlib.import_module("bokeh.models")
bokeh_plotting = importlib.import_module("bokeh.plotting")
import os
import time
import numpy as np
from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from sklearn.datasets import make_friedman1
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.inspection import partial_dependence
# Use imported modules
output_file = bokeh_io.output_file
save = bokeh_io.save
Band = bokeh_models.Band
ColumnDataSource = bokeh_models.ColumnDataSource
Span = bokeh_models.Span
figure = bokeh_plotting.figure
# 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" # Okabe-Ito position 1
ACCENT = "#C475FD" # Okabe-Ito position 2
# Data - Train a model and compute partial dependence
np.random.seed(42)
# Use Friedman #1 dataset which has known non-linear relationships
X, y = make_friedman1(n_samples=500, n_features=5, noise=0.5, random_state=42)
# 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 (has sin relationship)
feature_idx = 0
grid_resolution = 100
# Compute partial dependence using sklearn
pdp_results = partial_dependence(model, X, features=[feature_idx], kind="both", grid_resolution=grid_resolution)
# Extract values
avg_predictions = pdp_results["average"][0]
individual_predictions = pdp_results["individual"][0] # ICE lines
grid_values = pdp_results["grid_values"][0]
# Calculate confidence interval (percentiles of ICE lines)
lower_bound = np.percentile(individual_predictions, 10, axis=0)
upper_bound = np.percentile(individual_predictions, 90, axis=0)
# Center partial dependence at zero for easier interpretation
center_val = avg_predictions.mean()
avg_centered = avg_predictions - center_val
lower_centered = lower_bound - center_val
upper_centered = upper_bound - center_val
# Get training data distribution for rug plot
rug_x = X[:, feature_idx]
# Create data source for main line and band
source = ColumnDataSource(data={"x": grid_values, "y": avg_centered, "lower": lower_centered, "upper": upper_centered})
# Create data source for rug plot - position at bottom of plot area
y_min = lower_centered.min() - 1.5
rug_source = ColumnDataSource(data={"x": rug_x, "y": np.full_like(rug_x, y_min + 0.3)})
# Plot
p = figure(
width=4800,
height=2700,
title="pdp-basic · bokeh · anyplot.ai",
x_axis_label="Feature X₀ Value",
y_axis_label="Partial Dependence (centered)",
)
# Add confidence band
band = Band(
base="x",
lower="lower",
upper="upper",
source=source,
fill_color=BRAND,
fill_alpha=0.25,
line_color=BRAND,
line_alpha=0.4,
)
p.add_layout(band)
# Add horizontal line at y=0 for reference
zero_line = Span(location=0, dimension="width", line_color=INK_SOFT, line_width=3, line_dash="dashed", line_alpha=0.6)
p.add_layout(zero_line)
# Add invisible patch for confidence band legend entry
p.patch([], [], fill_color=BRAND, fill_alpha=0.25, line_color=BRAND, line_alpha=0.4, legend_label="80% CI")
# Add main PDP line
p.line("x", "y", source=source, line_width=5, line_color=BRAND, legend_label="Average PD")
# Add rug plot for data distribution
p.scatter(
"x",
"y",
source=rug_source,
size=25,
color=ACCENT,
alpha=0.6,
line_width=3,
angle=1.5708,
marker="dash",
legend_label="Data Distribution",
)
# Style - Text sizing
p.title.text_font_size = "28pt"
p.title.text_color = INK
p.xaxis.axis_label_text_font_size = "22pt"
p.yaxis.axis_label_text_font_size = "22pt"
p.xaxis.axis_label_text_color = INK
p.yaxis.axis_label_text_color = INK
p.xaxis.major_label_text_font_size = "18pt"
p.yaxis.major_label_text_font_size = "18pt"
p.xaxis.major_label_text_color = INK_SOFT
p.yaxis.major_label_text_color = INK_SOFT
# Axis styling
p.xaxis.axis_line_width = 3
p.yaxis.axis_line_width = 3
p.xaxis.axis_line_color = INK_SOFT
p.yaxis.axis_line_color = INK_SOFT
p.xaxis.major_tick_line_width = 3
p.yaxis.major_tick_line_width = 3
p.xaxis.major_tick_line_color = INK_SOFT
p.yaxis.major_tick_line_color = INK_SOFT
p.xaxis.minor_tick_line_width = 2
p.yaxis.minor_tick_line_width = 2
p.xaxis.minor_tick_line_color = INK_SOFT
p.yaxis.minor_tick_line_color = INK_SOFT
# Grid styling
p.xgrid.grid_line_color = INK
p.ygrid.grid_line_color = INK
p.xgrid.grid_line_alpha = 0.10
p.ygrid.grid_line_alpha = 0.10
p.xgrid.grid_line_dash = "dashed"
p.ygrid.grid_line_dash = "dashed"
# Legend styling
p.legend.location = "bottom_right"
p.legend.label_text_font_size = "18pt"
p.legend.label_text_color = INK_SOFT
p.legend.background_fill_color = ELEVATED_BG
p.legend.background_fill_alpha = 0.9
p.legend.border_line_color = INK_SOFT
p.legend.border_line_alpha = 0.5
p.legend.border_line_width = 2
p.legend.glyph_height = 50
p.legend.glyph_width = 50
p.legend.spacing = 20
p.legend.padding = 25
p.legend.margin = 40
# Background and border
p.background_fill_color = PAGE_BG
p.border_fill_color = PAGE_BG
p.outline_line_color = INK_SOFT
# Save HTML
output_file(f"plot-{THEME}.html")
save(p)
# Screenshot with headless Chrome
W, H = 4800, 2700
opts = Options()
for arg in (
"--headless=new",
"--no-sandbox",
"--disable-dev-shm-usage",
"--disable-gpu",
f"--window-size={W},{H}",
"--hide-scrollbars",
):
opts.add_argument(arg)
driver = webdriver.Chrome(options=opts)
driver.set_window_size(W, H)
driver.get(f"file://{Path(f'plot-{THEME}.html').resolve()}")
time.sleep(3) # let bokeh's JS render the canvas
driver.save_screenshot(f"plot-{THEME}.png")
driver.quit()
Part of Partial Dependence Plot on anyplot.ai.