Lesson 12

Horizontal Bar Charts

Use plt.barh when category names are long, set the bar thickness with height, and order the bars so the chart reads from the top down.

Intermediate25 min

Horizontal Bar Charts hero Module 25 Lesson 12

Turn a bar chart on its side and the category names get a whole line each instead of a cramped slot under a bar. That is the one real reason to prefer barh: long labels, or many categories. The numbers still run along the value axis, which is now the x axis.

A horizontal bar chart of deliveries by region

Same data, turned sideways

plt.barh(y, width) takes the categories first and the numbers second, exactly like bar. What changes is which axis carries which: the categories run down the y axis and the values along the x axis, so xlabel now names the measurement.

python

import matplotlib.pyplot as plt

regions = ["Northgate", "Eastport", "Westfield", "Southbank"]
deliveries = [148, 96, 172, 63]

plt.barh(regions, deliveries, color="#2563EB")
plt.title("Deliveries by region")
plt.xlabel("deliveries")
plt.show()

print("busiest:", regions[deliveries.index(max(deliveries))])

Output:

text

busiest: Westfield

Thickness with height

bar has width; barh has height. Both mean the thickness of the bar in category slots, and both default to 0.8. Thinner bars make a crowded chart breathe.

python

import matplotlib.pyplot as plt

regions = ["Northgate", "Eastport", "Westfield", "Southbank"]
deliveries = [148, 96, 172, 63]

plt.barh(regions, deliveries, height=0.5, color="#7C3AED")
plt.title("Thinner bars, more white space")
plt.xlabel("deliveries")
plt.show()

print("bar takes width, barh takes height")

Output:

text

bar takes width, barh takes height

Reading order from the top

Sorted data on a horizontal chart comes out upside down, because the y axis counts upwards from the bottom. Sort ascending and the largest bar lands at the top, which is where a reader starts. gca().invert_yaxis() is the other route to the same place.

python

import matplotlib.pyplot as plt

regions = ["Northgate", "Eastport", "Westfield", "Southbank"]
deliveries = [148, 96, 172, 63]

ordered = sorted(zip(deliveries, regions))
widths = [value for value, _ in ordered]
labels = [name for _, name in ordered]

plt.barh(labels, widths, color="#7C3AED")
plt.title("Largest at the top")
plt.xlabel("deliveries")
plt.show()

print("drawn bottom to top:", ", ".join(labels))

Output:

text

drawn bottom to top: Southbank, Eastport, Northgate, Westfield

Writing the value on the bar

When the exact number matters, print it at the end of the bar with plt.text. It takes an x position, a y position, and the string, and va="center" lines the text up with the middle of the bar.

python

import matplotlib.pyplot as plt

regions = ["Northgate", "Eastport", "Westfield", "Southbank"]
deliveries = [148, 96, 172, 63]

plt.barh(regions, deliveries, color="#7C3AED")
for position, value in enumerate(deliveries):
    plt.text(value + 3, position, str(value), va="center")

plt.title("Deliveries by region")
plt.xlabel("deliveries")
plt.xlim(0, max(deliveries) * 1.15)
plt.show()

print("the extra xlim leaves room for the labels")

Output:

text

the extra xlim leaves room for the labels

Key takeaways

barh puts categories on the y axis and values on the x axis
Reach for it when the category names are long or numerous
bar takes width, barh takes height, both in category slots
Sort ascending so the largest bar ends up at the top
Next: Stacked Bar Charts
Previous: Bar Charts

Loading playground…

Check your understanding

Horizontal Bar Charts Quiz

When is barh the better choice?

1 / 10

← Bar Charts

Stacked Bar Charts →