Lesson 20

Styling a Pie Chart

Pull one slice out with explode, rotate the pie with startangle, add a shadow or a donut hole, and put the labels in a legend instead.

Intermediate25 min

Styling a Pie Chart hero Module 25 Lesson 20

A plain pie gives every slice equal attention. The styling options let you point at the one slice that matters, turn the pie so it reads from the top, and move crowded labels out of the way.

A pie chart with the pandas slice pulled out from the centre

Pull a slice out with explode

explode takes one offset per slice, as a fraction of the radius. Zero leaves a slice in place and a small value such as 0.1 slides it outwards. Explode one slice, not all of them, or nothing stands out.

python

import matplotlib.pyplot as plt

tracks = ["Python", "NumPy", "pandas", "Plotting"]
enrolled = [120, 86, 94, 58]
explode = [0, 0, 0.12, 0]

plt.pie(enrolled, labels=tracks, explode=explode, autopct="%1.0f%%")
plt.title("pandas, pulled forward")
plt.show()

print(dict(zip(tracks, explode)))

Output:

text

{'Python': 0, 'NumPy': 0, 'pandas': 0.12, 'Plotting': 0}

Rotate with startangle

startangle turns the whole pie, in degrees counter-clockwise from three o'clock. startangle=90 starts the first slice at twelve o'clock, which is how most people expect to read a pie. counterclock=False makes the slices run clockwise instead.

python

import matplotlib.pyplot as plt

tracks = ["Python", "NumPy", "pandas", "Plotting"]
enrolled = [120, 86, 94, 58]

plt.pie(enrolled, labels=tracks, startangle=90, counterclock=False,
        autopct="%1.0f%%")
plt.title("Reading clockwise from the top")
plt.show()

print("start at 90 degrees")

Output:

text

start at 90 degrees

A donut with wedgeprops

wedgeprops passes styling to every slice. Setting a width smaller than 1 leaves a hole in the middle, which gives you room for a total. edgecolor draws a thin gap between slices.

python

import matplotlib.pyplot as plt

tracks = ["Python", "NumPy", "pandas", "Plotting"]
enrolled = [120, 86, 94, 58]

plt.pie(enrolled, labels=tracks, startangle=90,
        wedgeprops={"width": 0.4, "edgecolor": "white"})
plt.text(0, 0, str(sum(enrolled)), ha="center", va="center", fontsize=18)
plt.title("A donut with the total in the middle")
plt.show()

print("total in the hole:", sum(enrolled))

Output:

text

total in the hole: 358

Labels in a legend

When labels crowd each other, leave them off the slices and draw a legend. pie returns the wedges it drew, and plt.legend pairs them with your names. bbox_to_anchor moves the legend beside the pie so it covers nothing.

python

import matplotlib.pyplot as plt

tracks = ["Python", "NumPy", "pandas", "Plotting"]
enrolled = [120, 86, 94, 58]

wedges, texts, autotexts = plt.pie(enrolled, autopct="%1.0f%%", startangle=90)
plt.legend(wedges, tracks, title="track", loc="center left",
           bbox_to_anchor=(1, 0.5))
plt.title("Legend beside the pie")
plt.show()

print(len(wedges), "wedges")

Output:

text

4 wedges

With autopct set, pie returns three lists: the wedges, the label texts, and the percentage texts. Without it, only the first two.

A shadow, used sparingly

shadow=True adds a soft shadow under the pie. It looks dramatic but adds nothing to the data, so keep it for slides rather than reports.

python

import matplotlib.pyplot as plt

tracks = ["Python", "NumPy", "pandas", "Plotting"]
enrolled = [120, 86, 94, 58]

plt.pie(enrolled, labels=tracks, shadow=True, startangle=90,
        explode=[0.08, 0, 0, 0], autopct="%1.0f%%")
plt.title("Python leads, with a shadow")
plt.show()

print("shadow on")

Output:

text

shadow on

Key takeaways

explode holds one offset per slice; pull out one slice, not all
startangle=90 starts at twelve o'clock; counterclock=False runs clockwise
wedgeprops={"width": 0.4} turns a pie into a donut
Move crowded labels into a legend using the wedges pie returns
Next: Saving Figures
Previous: Pie Charts

Loading playground…

Check your understanding

Styling a Pie Chart Quiz

What does explode do?

1 / 10

← Pie Charts

Saving Figures →