A residual plot displays the difference between observed and predicted values (residuals) against fitted values or an independent variable in regression analysis. This diagnostic visualization helps identify violations of regression assumptions including non-linearity, heteroscedasticity (non-constant variance), and outliers. A well-fitting model shows residuals randomly scattered around zero with no discernible pattern.
""" anyplot.ai
residual-plot: Residual Plot
Library: pygal 3.1.0 | Python 3.13.13
Quality: 95/100 | Updated: 2026-05-10
"""
import os
import numpy as np
import pygal
from pygal.style import Style
# Theme tokens
THEME = os.getenv("ANYPLOT_THEME", "light")
PAGE_BG = "#FAF8F1" if THEME == "light" else "#1A1A17"
INK = "#1A1A17" if THEME == "light" else "#F0EFE8"
INK_MUTED = "#6B6A63" if THEME == "light" else "#A8A79F"
# Okabe-Ito palette
IMPRINT = ("#009E73", "#C475FD", "#4467A3", "#BD8233", "#AE3030", "#2ABCCD", "#954477")
# Data - Linear regression example with heteroscedasticity
np.random.seed(42)
n_points = 100
# Generate fitted values (x-axis) - house price predictions in $1000s
fitted_values = np.linspace(150, 500, n_points)
# Generate residuals with heteroscedasticity and outliers
base_residuals = np.random.normal(0, 20, n_points)
heteroscedasticity = (fitted_values / 500) * np.random.normal(0, 15, n_points)
residuals = base_residuals + heteroscedasticity
# Add outliers
outlier_indices = [15, 45, 78]
residuals[outlier_indices] = [85, -75, 90]
# Calculate standard deviation for reference bands
std_residuals = np.std(residuals)
upper_band = 2 * std_residuals
lower_band = -2 * std_residuals
# Identify outliers (beyond 2 standard deviations)
outlier_mask = np.abs(residuals) > 2 * std_residuals
# Custom style for 4800x2700 canvas with theme-adaptive tokens
custom_style = Style(
background=PAGE_BG,
plot_background=PAGE_BG,
foreground=INK,
foreground_strong=INK,
foreground_subtle=INK_MUTED,
colors=IMPRINT,
title_font_size=28,
label_font_size=22,
major_label_font_size=20,
legend_font_size=18,
value_font_size=14,
stroke_width=3,
guide_stroke_color=INK_MUTED,
guide_stroke_dasharray="3, 3",
)
# Create XY scatter chart for residual plot
chart = pygal.XY(
width=4800,
height=2700,
style=custom_style,
title="residual-plot · pygal · anyplot.ai",
x_title="Fitted Values - Predicted Price ($1000s)",
y_title="Residuals - Actual minus Predicted ($1000s)",
show_legend=True,
legend_at_bottom=True,
legend_at_bottom_columns=5,
show_x_guides=True,
show_y_guides=True,
stroke=False,
dots_size=12,
truncate_legend=-1,
x_label_rotation=0,
xrange=(140, 510),
range=(-100, 110),
)
# Set explicit x-axis labels
chart.x_labels = [150, 200, 250, 300, 350, 400, 450, 500]
# Prepare data points - separate normal and outlier points
normal_points = [(float(fitted_values[i]), float(residuals[i])) for i in range(n_points) if not outlier_mask[i]]
outlier_points = [(float(fitted_values[i]), float(residuals[i])) for i in range(n_points) if outlier_mask[i]]
# Add data series
chart.add("Residuals", normal_points)
chart.add("Outliers (>2σ)", outlier_points)
# Add zero reference line
zero_line_points = [(float(x), 0.0) for x in np.linspace(150, 500, 50)]
chart.add("Zero Reference (Perfect Fit)", zero_line_points, stroke=True, show_dots=False, stroke_style={"width": 5})
# Add +2σ reference band line
upper_band_points = [(float(x), float(upper_band)) for x in np.linspace(150, 500, 50)]
chart.add(
"+2σ Threshold", upper_band_points, stroke=True, show_dots=False, stroke_style={"width": 3, "dasharray": "10, 8"}
)
# Add -2σ reference band line
lower_band_points = [(float(x), float(lower_band)) for x in np.linspace(150, 500, 50)]
chart.add(
"-2σ Threshold", lower_band_points, stroke=True, show_dots=False, stroke_style={"width": 3, "dasharray": "10, 8"}
)
# Save as PNG and HTML
chart.render_to_png(f"plot-{THEME}.png")
with open(f"plot-{THEME}.html", "wb") as f:
f.write(chart.render())
Part of Residual Plot on anyplot.ai.