Horizontal grouped bars in Matplotlib - Deep Dive

Overview - Horizontal grouped bars

What is it?

Horizontal grouped bars are a type of bar chart where bars are drawn horizontally and grouped side-by-side to compare multiple categories across different groups. Each group contains bars representing different sub-categories, aligned horizontally for easy comparison. This visualization helps show relationships and differences clearly when categories have multiple sub-groups. It is especially useful when category names are long or when comparing multiple series side by side.

Why it matters

Without horizontal grouped bars, comparing multiple categories with sub-groups can be confusing or cluttered, especially if category labels are long or if vertical space is limited. This chart type makes it easier to see differences and patterns across groups at a glance, improving decision-making and communication. It helps businesses, researchers, and analysts quickly understand complex data with multiple dimensions.

Where it fits

Before learning horizontal grouped bars, you should understand basic bar charts and how to plot simple horizontal bars in matplotlib. After mastering this, you can explore stacked bars, grouped bars with vertical orientation, and advanced customization like annotations and interactive plots.

Mental Model

Core Idea

Horizontal grouped bars arrange multiple related bars side-by-side along the horizontal axis to compare sub-categories within each main category clearly.

Think of it like...

Imagine a row of bookshelves where each shelf represents a category, and on each shelf, different colored books represent sub-categories placed side by side for easy comparison.

Categories ──────────────▶
┌─────────────────────────────┐
│ Group 1: ███ ████ ██        │
│ Group 2: █████ ██ ███       │
│ Group 3: ██ ██████ █        │
└─────────────────────────────┘
Bars run horizontally, grouped by category rows.

Build-Up - 6 Steps

1

Foundation

Basic horizontal bar chart

Concept: Learn how to create a simple horizontal bar chart using matplotlib.

import matplotlib.pyplot as plt categories = ['A', 'B', 'C'] values = [5, 7, 3] plt.barh(categories, values) plt.xlabel('Value') plt.ylabel('Category') plt.title('Simple Horizontal Bar Chart') plt.show()

Result

A horizontal bar chart with three bars labeled A, B, and C, showing values 5, 7, and 3 respectively.

Understanding how to plot horizontal bars is the foundation for creating grouped horizontal bars later.

2

Foundation

Understanding bar positioning

Concept: Learn how bar positions are controlled on the axis and how to adjust them.

import matplotlib.pyplot as plt import numpy as np categories = ['A', 'B', 'C'] values1 = [5, 7, 3] values2 = [6, 2, 4] indices = np.arange(len(categories)) plt.barh(indices, values1, height=0.4, label='Group 1') plt.barh(indices + 0.4, values2, height=0.4, label='Group 2') plt.yticks(indices + 0.2, categories) plt.xlabel('Value') plt.title('Bar Positioning Example') plt.legend() plt.show()

Result

Two sets of horizontal bars appear side by side for each category, showing how shifting positions controls grouping.

Knowing how to shift bar positions along the axis is key to grouping bars horizontally.

3

Intermediate

Creating horizontal grouped bars

🤔Before reading on: do you think you can group bars horizontally by shifting their positions on the y-axis or x-axis? Commit to your answer.

Concept: Combine multiple horizontal bars per category by adjusting their vertical positions to create groups.

import matplotlib.pyplot as plt import numpy as np categories = ['A', 'B', 'C'] values1 = [5, 7, 3] values2 = [6, 2, 4] values3 = [4, 5, 6] indices = np.arange(len(categories)) bar_height = 0.2 plt.barh(indices - bar_height, values1, height=bar_height, label='Group 1') plt.barh(indices, values2, height=bar_height, label='Group 2') plt.barh(indices + bar_height, values3, height=bar_height, label='Group 3') plt.yticks(indices, categories) plt.xlabel('Value') plt.title('Horizontal Grouped Bars') plt.legend() plt.show()

Result

A horizontal grouped bar chart with three bars per category, grouped side by side horizontally.

Shifting bars vertically by small amounts creates clear groups, enabling comparison of multiple series per category.

4

Intermediate

Customizing colors and labels

🤔Before reading on: do you think adding colors and labels improves clarity or just adds decoration? Commit to your answer.

Concept: Use colors and labels to make groups distinct and easier to interpret.

import matplotlib.pyplot as plt import numpy as np categories = ['A', 'B', 'C'] values1 = [5, 7, 3] values2 = [6, 2, 4] values3 = [4, 5, 6] indices = np.arange(len(categories)) bar_height = 0.2 colors = ['#1f77b4', '#ff7f0e', '#2ca02c'] plt.barh(indices - bar_height, values1, height=bar_height, color=colors[0], label='Group 1') plt.barh(indices, values2, height=bar_height, color=colors[1], label='Group 2') plt.barh(indices + bar_height, values3, height=bar_height, color=colors[2], label='Group 3') plt.yticks(indices, categories) plt.xlabel('Value') plt.title('Colored Horizontal Grouped Bars') plt.legend() plt.show()

Result

Bars are colored distinctly per group, with a legend explaining each color, improving readability.

Colors and legends help viewers quickly identify groups and understand the chart without confusion.

5

Advanced

Adding value labels on bars

🤔Before reading on: do you think adding numeric labels on bars makes the chart cluttered or more informative? Commit to your answer.

Concept: Display numeric values on each bar to provide exact data points alongside visual comparison.

import matplotlib.pyplot as plt import numpy as np categories = ['A', 'B', 'C'] values1 = [5, 7, 3] values2 = [6, 2, 4] values3 = [4, 5, 6] indices = np.arange(len(categories)) bar_height = 0.2 colors = ['#1f77b4', '#ff7f0e', '#2ca02c'] bars1 = plt.barh(indices - bar_height, values1, height=bar_height, color=colors[0], label='Group 1') bars2 = plt.barh(indices, values2, height=bar_height, color=colors[1], label='Group 2') bars3 = plt.barh(indices + bar_height, values3, height=bar_height, color=colors[2], label='Group 3') plt.yticks(indices, categories) plt.xlabel('Value') plt.title('Horizontal Grouped Bars with Labels') plt.legend() for bars in [bars1, bars2, bars3]: for bar in bars: width = bar.get_width() plt.text(width + 0.1, bar.get_y() + bar.get_height()/2, f'{width}', va='center') plt.show()

Result

Each bar shows its numeric value at the end, making exact comparisons easy.

Adding labels balances visual and numeric information, aiding precise interpretation.

6

Expert

Handling uneven group sizes and spacing

🤔Before reading on: do you think grouped bars must always have the same number of bars per group? Commit to your answer.

Concept: Learn how to plot grouped bars when groups have different numbers of sub-categories and how to adjust spacing for clarity.

import matplotlib.pyplot as plt import numpy as np categories = ['A', 'B', 'C'] values = { 'Group 1': [5, 7], 'Group 2': [6, 2, 4], 'Group 3': [4] } max_bars = max(len(v) for v in values.values()) indices = np.arange(len(categories)) bar_height = 0.15 for i, (group, vals) in enumerate(values.items()): positions = indices - (bar_height * (len(vals) - 1) / 2) + i * bar_height plt.barh(positions[:len(vals)], vals, height=bar_height, label=group) plt.yticks(indices, categories) plt.xlabel('Value') plt.title('Grouped Bars with Uneven Sizes') plt.legend() plt.show()

Result

Bars are grouped horizontally with different numbers of bars per category, spaced evenly for readability.

Handling uneven group sizes requires dynamic positioning to maintain clear grouping and avoid overlap.

Under the Hood

Matplotlib draws horizontal bars by plotting rectangles along the y-axis positions with widths representing values. Grouping bars horizontally involves calculating y-axis positions for each bar within a category, offsetting them by a fixed height to avoid overlap. The bar height controls thickness, and the y-ticks are adjusted to label the center of each category group. Internally, matplotlib uses patches.Rectangle objects for bars, and the rendering engine draws them in order.

Why designed this way?

This design allows flexible control over bar placement and size, enabling grouped bars by simply shifting positions. It avoids complex layout engines by relying on numeric axis positioning. Alternatives like stacked bars overlay bars vertically, but grouping side-by-side requires explicit position control. This approach balances simplicity, performance, and customization.

Categories (y-axis)
┌─────────────────────────────┐
│  ┌─────┐  ┌──────┐  ┌───┐   │
│  │Bar1 │  │Bar2  │  │Bar3│   │
│  └─────┘  └──────┘  └───┘   │
│  Positions shifted vertically│
│  Bars drawn as rectangles    │
└─────────────────────────────┘
Values control bar width (x-axis)

Myth Busters - 4 Common Misconceptions

Quick: Do horizontal grouped bars require bars to be stacked vertically? Commit to yes or no.

Common Belief: Grouped bars must be stacked on top of each other vertically to show groups.

Reality: Grouped bars are placed side by side horizontally by shifting their vertical positions, not stacked on top.

Why it matters: Stacking bars instead of grouping side by side changes the visual meaning and can confuse comparisons between groups.

Quick: Do you think bar height controls the length of the bar in horizontal bar charts? Commit to yes or no.

Common Belief: Bar height controls how long the bar is in horizontal bar charts.

Reality: Bar height controls the thickness of the bar; the length is controlled by the value plotted along the x-axis.

Why it matters: Misunderstanding this leads to incorrect bar sizing and misleading visualizations.

Quick: Can you group bars horizontally without adjusting their y-axis positions? Commit to yes or no.

Common Belief: Bars automatically group horizontally without manual position adjustments.

Reality: You must manually shift bars' y-axis positions to create horizontal groups; matplotlib does not auto-group them.

Why it matters: Assuming automatic grouping causes overlapping bars and unreadable charts.

Quick: Do you think horizontal grouped bars are less readable than vertical ones? Commit to yes or no.

Common Belief: Horizontal grouped bars are harder to read than vertical grouped bars.

Reality: Horizontal grouped bars are often more readable, especially when category labels are long or numerous.

Why it matters: Choosing vertical bars by default may reduce clarity when horizontal grouping is better suited.

Expert Zone

1

Precise control of bar height and spacing is crucial to avoid visual clutter or overlap, especially with many groups.

2

When groups have uneven numbers of bars, dynamic calculation of positions prevents misalignment and maintains readability.

3

Using alpha transparency and hatch patterns can help distinguish groups when color palettes are limited or for colorblind accessibility.

When NOT to use

Avoid horizontal grouped bars when you have too many groups or sub-categories, as the chart becomes cluttered and hard to read. Instead, consider interactive plots, small multiples, or summary statistics. Also, if the focus is on total composition rather than comparison, stacked bars or pie charts might be better.

Production Patterns

Professionals use horizontal grouped bars in dashboards to compare sales across regions and product lines, in reports to show survey results by demographic groups, and in presentations to highlight performance differences. They often combine these charts with annotations, tooltips, and consistent color schemes for clarity.

Connections

Stacked bar charts

Alternative way to show multiple sub-categories within categories by stacking bars vertically instead of grouping side by side.

Understanding grouped bars clarifies when stacking is better for showing part-to-whole relationships versus side-by-side comparison.

Data visualization principles

Horizontal grouped bars apply core principles like pre-attentive processing and spatial grouping to improve data comprehension.

Knowing visualization principles helps design grouped bars that communicate data clearly and avoid misleading impressions.

Music notation (different instruments in a score)

Like grouped bars, music scores group different instruments' notes horizontally aligned by time, allowing simultaneous comparison.

Recognizing grouping patterns across fields shows how arranging related elements side by side aids understanding complex information.

Common Pitfalls

#1 Bars overlap because positions are not shifted correctly.

Wrong approach:

plt.barh(indices, values1, height=0.4)
plt.barh(indices, values2, height=0.4)

Correct approach:

plt.barh(indices - 0.2, values1, height=0.4)
plt.barh(indices + 0.2, values2, height=0.4)

Root cause:

Not offsetting y-axis positions causes bars to draw on top of each other.

#2 Using the same color for all groups, causing confusion.

Wrong approach:

plt.barh(indices - 0.2, values1, height=0.4, color='blue')
plt.barh(indices + 0.2, values2, height=0.4, color='blue')

Correct approach:

plt.barh(indices - 0.2, values1, height=0.4, color='blue')
plt.barh(indices + 0.2, values2, height=0.4, color='orange')

Root cause:

Not differentiating groups visually makes it hard to distinguish bars.

#3 Placing y-ticks incorrectly, misaligning labels with bars.

Wrong approach:

plt.yticks(indices + 0.2, categories)

Correct approach:

plt.yticks(indices, categories)

Root cause:

Misunderstanding that y-ticks should align with the center of grouped bars, not shifted positions.

Key Takeaways

Horizontal grouped bars display multiple sub-categories side by side along the horizontal axis for clear comparison.

Shifting bar positions vertically is essential to create distinct groups and avoid overlap.

Colors, labels, and legends improve readability and help viewers quickly identify groups.

Handling uneven group sizes requires dynamic positioning to maintain clarity.

Understanding the internal mechanics of bar placement helps avoid common mistakes and design effective visualizations.