Permutation Feature Importance Plot — lets-plot

A horizontal bar chart displaying permutation feature importance from machine learning models, showing the decrease in model score when each feature is randomly shuffled. Unlike model-specific feature importances, permutation importance is model-agnostic and measures how much the model's performance degrades when a feature's relationship with the target is broken. Error bars indicate variability across multiple shuffles, providing a confidence measure for each importance score.

Permutation Feature Importance Plot rendered with lets-plot

Python source (lets-plot)

""" anyplot.ai
bar-permutation-importance: Permutation Feature Importance Plot
Library: letsplot 4.9.0 | Python 3.13.13
Quality: 84/100 | Updated: 2026-05-17
"""

import os

import numpy as np
import pandas as pd
from lets_plot import (
    LetsPlot,
    aes,
    coord_flip,
    element_line,
    element_rect,
    element_text,
    geom_bar,
    geom_errorbar,
    geom_hline,
    geom_text,
    ggplot,
    ggsave,
    ggsize,
    labs,
    scale_fill_gradient,
    scale_x_discrete,
    theme,
    theme_minimal,
)


LetsPlot.setup_html()

# 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"
INK_MUTED = "#8A8A7E" if THEME == "light" else "#7A7970"

# Data: Simulated permutation importance from a Random Forest model
np.random.seed(42)

features = [
    "Income Level",
    "Credit Score",
    "Employment Years",
    "Debt Ratio",
    "Account Age",
    "Payment History",
    "Loan Amount",
    "Interest Rate",
    "Property Value",
    "Monthly Expenses",
    "Savings Balance",
    "Number of Accounts",
    "Recent Inquiries",
    "Education Level",
    "Region Code",
]

# Generate importance values - higher for more predictive features
base_importance = np.array(
    [0.15, 0.12, 0.09, 0.08, 0.06, 0.05, 0.04, 0.03, 0.025, 0.02, 0.015, 0.01, 0.008, 0.005, 0.002]
)
# Add some noise
importance_mean = base_importance + np.random.uniform(-0.005, 0.005, len(features))
importance_std = np.random.uniform(0.003, 0.02, len(features))

# Create DataFrame and sort by importance
df = pd.DataFrame({"feature": features, "importance_mean": importance_mean, "importance_std": importance_std})
df = df.sort_values("importance_mean", ascending=True).reset_index(drop=True)

# Create ordered categorical for proper y-axis ordering
df["feature"] = pd.Categorical(df["feature"], categories=df["feature"].tolist(), ordered=True)

# Calculate error bar positions and label format
df["ymin"] = df["importance_mean"] - df["importance_std"]
df["ymax"] = df["importance_mean"] + df["importance_std"]
df["importance_pct"] = (df["importance_mean"] * 100).round(1).astype(str) + "%"
df["label"] = df["importance_mean"].round(4).astype(str)

# Create the plot with enhanced letsplot-specific features
plot = (
    ggplot(df, aes(x="feature", y="importance_mean", fill="importance_mean"))
    + geom_bar(
        stat="identity",
        width=0.75,
        alpha=0.92,
        tooltip=aes(text="<b>@feature</b><br/>Importance: @label<br/>±Std: @importance_std|.0000"),
    )
    + geom_errorbar(aes(ymin="ymin", ymax="ymax"), width=0.3, size=1.0, color=INK_SOFT, alpha=0.8)
    + geom_text(aes(y="importance_mean", label="label"), vjust=-0.6, size=12, color=INK_MUTED, family="monospace")
    + geom_hline(yintercept=0, color=INK_SOFT, size=0.9, linetype="dashed", alpha=0.6)
    + coord_flip()
    + scale_fill_gradient(low="#FFD43B", high="#306998", guide="none")
    + scale_x_discrete()
    + labs(x="Feature", y="Mean Decrease in Model Score", title="bar-permutation-importance · letsplot · anyplot.ai")
    + theme_minimal()
    + theme(
        plot_background=element_rect(fill=PAGE_BG, color=PAGE_BG),
        panel_background=element_rect(fill=PAGE_BG, color=PAGE_BG),
        panel_grid_major=element_line(color=INK_SOFT, size=0.3),
        panel_grid_minor_x=element_line(color=INK_SOFT, size=0.15),
        plot_title=element_text(size=24, color=INK, face="bold"),
        axis_title=element_text(size=20, color=INK, face="bold"),
        axis_text=element_text(size=16, color=INK_SOFT),
        axis_text_y=element_text(size=14, color=INK_SOFT),
    )
    + ggsize(1600, 900)
)

# Save as PNG (scale 3x for 4800x2700)
ggsave(plot, f"plot-{THEME}.png", path=".", scale=3)

# Save as HTML for interactive view
ggsave(plot, f"plot-{THEME}.html", path=".")

Part of Permutation Feature Importance Plot on anyplot.ai.

Other implementations