A parallel categories plot visualizes categorical data across multiple dimensions, with vertical axes representing each categorical variable and ribbons connecting categories to show observation flow. Unlike parallel coordinates (which use lines for numeric data), parallel categories use width-proportional ribbons to show counts or frequencies, making it ideal for understanding how categorical values co-occur and flow across multiple classification dimensions.
""" anyplot.ai
parallel-categories-basic: Basic Parallel Categories Plot
Library: seaborn 0.13.2 | Python 3.13.13
Quality: 63/100 | Updated: 2026-05-13
"""
import matplotlib.patches as mpatches
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
from matplotlib.path import Path
# Set seaborn style for consistent aesthetics
sns.set_style("whitegrid")
sns.set_context("talk", font_scale=1.2)
# Use seaborn's colorblind-safe palette for accessibility
class_palette = sns.color_palette("colorblind", 3)
class_colors = {"First": class_palette[0], "Second": class_palette[1], "Third": class_palette[2]}
# Data - Titanic-style dataset with categorical dimensions
np.random.seed(42)
n_samples = 500
data = {
"Class": np.random.choice(["First", "Second", "Third"], n_samples, p=[0.25, 0.25, 0.50]),
"Sex": np.random.choice(["Male", "Female"], n_samples, p=[0.55, 0.45]),
"Age Group": np.random.choice(["Child", "Adult", "Senior"], n_samples, p=[0.15, 0.70, 0.15]),
"Embarked": np.random.choice(["Southampton", "Cherbourg", "Queenstown"], n_samples, p=[0.70, 0.20, 0.10]),
}
# Create survival based on realistic patterns
survival_prob = np.zeros(n_samples)
for i in range(n_samples):
p = 0.3
if data["Class"][i] == "First":
p += 0.35
elif data["Class"][i] == "Second":
p += 0.15
if data["Sex"][i] == "Female":
p += 0.25
if data["Age Group"][i] == "Child":
p += 0.15
survival_prob[i] = min(p, 0.95)
data["Outcome"] = np.where(np.random.random(n_samples) < survival_prob, "Survived", "Lost")
df = pd.DataFrame(data)
# Aggregate data for parallel categories
dimensions = ["Class", "Sex", "Age Group", "Embarked", "Outcome"]
dim_orders = {
"Class": ["First", "Second", "Third"],
"Sex": ["Female", "Male"],
"Age Group": ["Child", "Adult", "Senior"],
"Embarked": ["Southampton", "Cherbourg", "Queenstown"],
"Outcome": ["Survived", "Lost"],
}
# Create figure - removed inset to reduce crowding
fig, ax = plt.subplots(figsize=(16, 9))
# Calculate positions for each dimension
n_dims = len(dimensions)
x_positions = np.linspace(0.10, 0.90, n_dims)
dim_width = 0.025
# Track category positions and heights
category_positions = {}
# Draw category bars and labels
for dim_idx, dim in enumerate(dimensions):
x_pos = x_positions[dim_idx]
categories = dim_orders[dim]
counts = df[dim].value_counts()
total = counts.sum()
heights = {cat: counts.get(cat, 0) / total for cat in categories}
y_start, y_end = 0.08, 0.88
y_range = y_end - y_start
current_y = y_start
for cat in categories:
height = heights[cat] * y_range
category_positions[(dim, cat)] = (x_pos, current_y, height)
# Draw category rectangle
rect = mpatches.FancyBboxPatch(
(x_pos - dim_width / 2, current_y),
dim_width,
height,
boxstyle="round,pad=0.003,rounding_size=0.008",
facecolor="#3a3a3a",
edgecolor="white",
linewidth=1.5,
zorder=10,
)
ax.add_patch(rect)
# Place all category labels outside bars for better readability
label_y = current_y + height / 2
if dim_idx == 0:
# First dimension: labels on left
ax.text(
x_pos - dim_width / 2 - 0.015, label_y, cat, ha="right", va="center", fontsize=13, fontweight="bold"
)
elif dim_idx == n_dims - 1:
# Last dimension: labels on right
ax.text(x_pos + dim_width / 2 + 0.015, label_y, cat, ha="left", va="center", fontsize=13, fontweight="bold")
elif dim_idx == n_dims - 2:
# Embarked dimension: labels on right to avoid ribbon overlap
ax.text(x_pos + dim_width / 2 + 0.015, label_y, cat, ha="left", va="center", fontsize=11, fontweight="bold")
else:
# Other middle dimensions: labels on left
ax.text(
x_pos - dim_width / 2 - 0.012, label_y, cat, ha="right", va="center", fontsize=12, fontweight="bold"
)
current_y += height
# Draw ribbons connecting categories
for i in range(n_dims - 1):
dim1, dim2 = dimensions[i], dimensions[i + 1]
x1, x2 = x_positions[i], x_positions[i + 1]
flow_counts = df.groupby([dim1, dim2]).size().reset_index(name="count")
cat1_current = {cat: category_positions[(dim1, cat)][1] for cat in dim_orders[dim1]}
cat2_current = {cat: category_positions[(dim2, cat)][1] for cat in dim_orders[dim2]}
total_count = len(df)
y_range = 0.80
for _, row in flow_counts.iterrows():
cat1, cat2, count = row[dim1], row[dim2], row["count"]
ribbon_height = (count / total_count) * y_range
y1_bottom = cat1_current[cat1]
y2_bottom = cat2_current[cat2]
y1_top = y1_bottom + ribbon_height
y2_top = y2_bottom + ribbon_height
cat1_current[cat1] = y1_top
cat2_current[cat2] = y2_top
x_mid = (x1 + x2) / 2
verts = [
(x1 + dim_width / 2, y1_bottom),
(x_mid, y1_bottom),
(x_mid, y2_bottom),
(x2 - dim_width / 2, y2_bottom),
(x2 - dim_width / 2, y2_top),
(x_mid, y2_top),
(x_mid, y1_top),
(x1 + dim_width / 2, y1_top),
(x1 + dim_width / 2, y1_bottom),
]
codes = [
Path.MOVETO,
Path.CURVE3,
Path.CURVE3,
Path.LINETO,
Path.LINETO,
Path.CURVE3,
Path.CURVE3,
Path.LINETO,
Path.CLOSEPOLY,
]
path = Path(verts, codes)
# Color by Class category
first_cat = df.loc[(df[dim1] == cat1) & (df[dim2] == cat2), "Class"].mode()
color = class_colors.get(first_cat.iloc[0], class_palette[0]) if len(first_cat) > 0 else class_palette[0]
patch = mpatches.PathPatch(path, facecolor=color, edgecolor="white", linewidth=0.3, alpha=0.55, zorder=5)
ax.add_patch(patch)
# Add dimension labels at the top
for dim_idx, dim in enumerate(dimensions):
ax.text(
x_positions[dim_idx],
0.94,
dim,
ha="center",
va="bottom",
fontsize=17,
fontweight="bold",
color=class_palette[0],
)
# Legend for Class colors
legend_patches = [
mpatches.Patch(color=class_colors["First"], alpha=0.7, label="First Class"),
mpatches.Patch(color=class_colors["Second"], alpha=0.7, label="Second Class"),
mpatches.Patch(color=class_colors["Third"], alpha=0.7, label="Third Class"),
]
ax.legend(
handles=legend_patches,
loc="lower center",
fontsize=12,
framealpha=0.9,
edgecolor="gray",
ncol=3,
bbox_to_anchor=(0.5, -0.02),
)
# Style adjustments
ax.set_xlim(0, 1)
ax.set_ylim(0, 1.02)
ax.set_aspect("auto")
ax.axis("off")
# Title
ax.set_title("parallel-categories-basic · seaborn · pyplots.ai", fontsize=24, fontweight="bold", pad=15)
plt.savefig("plot.png", dpi=300, bbox_inches="tight", facecolor="white")
Part of Basic Parallel Categories Plot on anyplot.ai.