Violin Plots in Matplotlib
Learn how to create violin plots using Matplotlib. Understand data distribution, probability density, median, quartiles, bandwidth and violin plot customization.
1. Learning Objectives
After completing this lesson, you will be able to:
- Understand what a violin plot is.
- Understand probability density and data distribution.
- Create basic violin plots using Matplotlib.
- Compare multiple datasets using violin plots.
- Display means, medians and quartiles.
- Customize violin plot colours and appearance.
- Control violin width and bandwidth.
- Create horizontal violin plots.
- Understand the difference between box plots and violin plots.
2. Introduction to Violin Plots
A violin plot is a statistical chart used to visualize the distribution and density of numerical data. It combines information from a box plot and a probability density plot.
The width of a violin at a particular value represents the estimated density of observations around that value. Wider regions indicate a higher concentration of data, while narrower regions indicate a lower concentration.
Real-Life Example
Suppose a teacher wants to compare examination marks from three classes. A violin plot can show where most students scored, how widely marks are distributed and whether the data contains multiple concentration areas.
Conceptual illustration showing how violin width represents the density of observations.
3. Understanding Data Density
A violin plot uses kernel density estimation (KDE) to estimate the probability density of numerical observations. KDE creates a smooth curve representing where values are concentrated.
The estimated density is mirrored around a central position to create the violin shape. This is why the chart looks similar to a violin instrument.
Understanding Width
A wider part of the violin means many observations are concentrated around that range. A narrow part means relatively fewer observations are present.
Wide Region
= Higher estimated density
Narrow Region
= Lower estimated density
4. Import Matplotlib and NumPy
We use Matplotlib to create violin plots and NumPy to generate numerical sample data.
import matplotlib.pyplot as plt
import numpy as np
The plt.violinplot() function creates a violin
plot from one or more numerical datasets.
5. Creating a Basic Violin Plot
To create a basic violin plot, pass a numerical dataset
to plt.violinplot().
Example
import matplotlib.pyplot as plt
marks = [
45, 50, 52, 55, 58, 60,
62, 65, 68, 70, 72, 75
]
plt.violinplot(marks)
plt.title("Student Marks Distribution")
plt.ylabel("Marks")
plt.show()
Expected Output
Student Marks Distribution
A violin-shaped distribution is displayed.
The shape represents the estimated density
of student marks across the dataset.
6. Creating Multiple Violin Plots
You can compare multiple datasets by passing a list of
datasets to plt.violinplot().
import matplotlib.pyplot as plt
class_a = [45, 50, 55, 60, 65, 70, 75]
class_b = [50, 55, 60, 65, 70, 75, 80]
class_c = [40, 45, 50, 55, 60, 65, 70]
data = [class_a, class_b, class_c]
plt.violinplot(data)
plt.title("Comparison of Student Marks")
plt.ylabel("Marks")
plt.xticks(
[1, 2, 3],
["Class A", "Class B", "Class C"]
)
plt.show()
Expected Output
Comparison of Student Marks
Three violin plots are displayed:
Violin 1 = Class A
Violin 2 = Class B
Violin 3 = Class C
Each violin represents the distribution
of marks for its respective class.
Why use multiple violin plots?
Multiple violin plots help compare the distribution shape, density, spread and concentration of different datasets.
7. Adding Labels to Violin Plots
Use plt.xticks() to display category names
below each violin plot. This makes multiple datasets
easier to identify.
import matplotlib.pyplot as plt
group_a = [20, 25, 30, 35, 40, 45, 50]
group_b = [30, 35, 40, 45, 50, 55, 60]
data = [group_a, group_b]
plt.violinplot(data)
plt.xticks(
[1, 2],
["Group A", "Group B"]
)
plt.title("Group Distribution")
plt.ylabel("Values")
plt.show()
Expected Output
Group Distribution
Two violin plots are displayed.
Group A
Group B
Each violin is identified by its category label.
8. Displaying Mean and Median
Matplotlib provides the showmeans and
showmedians parameters to display statistical
indicators on violin plots.
import matplotlib.pyplot as plt
marks = [
45, 50, 52, 55, 58, 60,
62, 65, 68, 70, 72, 75
]
plt.violinplot(
marks,
showmeans=True,
showmedians=True
)
plt.title("Violin Plot with Mean and Median")
plt.ylabel("Marks")
plt.show()
Expected Output
Violin Plot with Mean and Median
A violin plot is displayed with:
- Estimated data distribution
- Mean indicator
- Median indicator
9. Displaying Minimum and Maximum Values
The showextrema parameter controls whether
the minimum, maximum and central range indicators are
displayed on the violin plot.
import matplotlib.pyplot as plt
marks = [45, 50, 55, 60, 65, 70, 75, 80]
plt.violinplot(
marks,
showextrema=True
)
plt.title("Violin Plot with Extrema")
plt.ylabel("Marks")
plt.show()
Expected Output
Violin Plot with Extrema
The violin plot displays:
- Minimum value
- Maximum value
- Central statistical indicator
Hiding Extrema
plt.violinplot(
marks,
showextrema=False
)
Expected Output
The violin shape is displayed without
the minimum, maximum and central range lines.
10. Displaying Quartiles with showquartiles
The showquartiles parameter controls whether
the first, second and third quartile indicators are
displayed.
import matplotlib.pyplot as plt
marks = [
45, 50, 52, 55, 58, 60,
62, 65, 68, 70, 72, 75
]
plt.violinplot(
marks,
showquartiles=True
)
plt.title("Violin Plot with Quartiles")
plt.ylabel("Marks")
plt.show()
Expected Output
Violin Plot with Quartiles
A violin plot is displayed with:
- First quartile (Q1)
- Median (Q2)
- Third quartile (Q3)
11. Customizing Violin Plot Colours
The plt.violinplot() function returns a dictionary
of graphical components. You can use the returned
bodies collection to customize violin colours.
import matplotlib.pyplot as plt
marks = [45, 50, 55, 60, 65, 70, 75]
fig, ax = plt.subplots()
result = ax.violinplot(marks)
for body in result["bodies"]:
body.set_facecolor("#93c5fd")
body.set_edgecolor("#1d4ed8")
body.set_alpha(0.8)
ax.set_title("Colored Violin Plot")
ax.set_ylabel("Marks")
plt.show()
Expected Output
Colored Violin Plot
The violin body is filled with light blue.
The violin border is dark blue.
The transparency is set to 0.8.
Common Styling Methods
| Method | Purpose |
|---|---|
set_facecolor() |
Changes the violin fill colour. |
set_edgecolor() |
Changes the violin border colour. |
set_alpha() |
Controls transparency. |
set_linewidth() |
Changes border line thickness. |
12. Applying Different Colours to Multiple Violins
You can assign different colours to individual violin bodies using a loop.
import matplotlib.pyplot as plt
class_a = [45, 50, 55, 60, 65, 70, 75]
class_b = [50, 55, 60, 65, 70, 75, 80]
class_c = [40, 45, 50, 55, 60, 65, 70]
data = [class_a, class_b, class_c]
fig, ax = plt.subplots()
result = ax.violinplot(data)
colors = [
"#93c5fd",
"#86efac",
"#fcd34d"
]
for body, color in zip(result["bodies"], colors):
body.set_facecolor(color)
body.set_edgecolor("#334155")
body.set_alpha(0.85)
ax.set_xticks(
[1, 2, 3],
["Class A", "Class B", "Class C"]
)
ax.set_title("Colored Class Comparison")
ax.set_ylabel("Marks")
plt.show()
Expected Output
Colored Class Comparison
Class A = Light Blue
Class B = Light Green
Class C = Light Yellow
Three differently coloured violin plots are displayed.
13. Creating Horizontal Violin Plots
By default, violin plots are vertical. You can create
horizontal violin plots using the
orientation="horizontal" parameter.
import matplotlib.pyplot as plt
marks = [45, 50, 55, 60, 65, 70, 75, 80]
plt.violinplot(
marks,
orientation="horizontal"
)
plt.title("Horizontal Violin Plot")
plt.xlabel("Marks")
plt.show()
Expected Output
Horizontal Violin Plot
A horizontal violin shape is displayed.
The numerical values are represented along
the X-axis.
14. Controlling Violin Width
The widths parameter controls the maximum
width of the violin bodies. It can be useful when
displaying multiple violin plots.
import matplotlib.pyplot as plt
data = [
[45, 50, 55, 60, 65, 70, 75],
[50, 55, 60, 65, 70, 75, 80],
[40, 45, 50, 55, 60, 65, 70]
]
plt.violinplot(
data,
widths=0.5
)
plt.title("Violin Plot with Custom Width")
plt.xticks(
[1, 2, 3],
["Group A", "Group B", "Group C"]
)
plt.ylabel("Values")
plt.show()
Expected Output
Violin Plot with Custom Width
Three violin plots are displayed.
Each violin uses a maximum width setting of 0.5.
15. Understanding Bandwidth with bw_method
The bw_method parameter controls the bandwidth
used by the kernel density estimation process. Bandwidth
influences how smooth or detailed the violin shape appears.
import matplotlib.pyplot as plt
import numpy as np
data = np.random.normal(
loc=50,
scale=10,
size=100
)
plt.violinplot(
data,
bw_method=0.5
)
plt.title("Violin Plot with Custom Bandwidth")
plt.ylabel("Values")
plt.show()
Expected Output
Violin Plot with Custom Bandwidth
A violin plot is displayed using a bandwidth
setting of 0.5.
The bandwidth affects the smoothness
of the estimated density shape.
Bandwidth Comparison
| Bandwidth | Effect |
|---|---|
| Smaller bandwidth | Can show more local detail and variation. |
| Larger bandwidth | Produces a smoother density estimate. |
16. Understanding Violin Plot Components
The violin plot function returns several graphical components that can be customized individually.
import matplotlib.pyplot as plt
marks = [45, 50, 55, 60, 65, 70, 75]
fig, ax = plt.subplots()
result = ax.violinplot(
marks,
showmeans=True,
showmedians=True,
showextrema=True
)
print(result.keys())
plt.show()
Expected Output
dict_keys([
'bodies',
'cmeans',
'cmaxes',
'cmins',
'cbars',
'cmedians'
])
The returned dictionary provides access to the violin bodies, mean indicators, maximum and minimum indicators, central range lines and median indicators. Exact available components depend on the selected options and Matplotlib version.
| Component | Description |
|---|---|
bodies |
The filled violin shapes. |
cmeans |
Mean indicator collection. |
cmedians |
Median indicator collection. |
cmins |
Minimum indicator collection. |
cmaxes |
Maximum indicator collection. |
cbars |
Central range line collection. |
17. Difference Between Box Plot and Violin Plot
Box plots and violin plots are both used to understand numerical data distributions. However, they display different levels of information.
| Box Plot | Violin Plot |
|---|---|
| Shows quartiles and median. | Shows estimated probability density. |
| Uses a rectangular box. | Uses a density-shaped body. |
| Summarizes the five-number distribution. | Shows the shape of the distribution. |
| Does not directly show density shape. | Shows where observations are concentrated. |
| Usually easier to interpret. | Provides more detail about distribution shape. |
When should you use each chart?
Use a box plot when you want a compact summary of the median, quartiles and possible outliers. Use a violin plot when you want to understand the shape and density of the distribution.
18. Violin Plots Using the Object-Oriented API
Matplotlib also supports an object-oriented approach.
Instead of calling plt.violinplot() directly,
create a Figure and Axes object and use
ax.violinplot().
import matplotlib.pyplot as plt
import numpy as np
data = [
np.random.normal(50, 10, 100),
np.random.normal(60, 12, 100),
np.random.normal(70, 8, 100)
]
fig, ax = plt.subplots(figsize=(8, 5))
result = ax.violinplot(
data,
showmeans=True,
showmedians=True
)
ax.set_xticks(
[1, 2, 3],
["Group A", "Group B", "Group C"]
)
ax.set_title("Group Distribution Comparison")
ax.set_ylabel("Values")
plt.show()
Expected Output
Group Distribution Comparison
Three violin plots are displayed.
Group A = Mean around 50
Group B = Mean around 60
Group C = Mean around 70
Each violin displays its estimated distribution,
mean indicator and median indicator.
19. Important Parameters of plt.violinplot()
| Parameter | Description | Example |
|---|---|---|
dataset |
Input numerical data. | plt.violinplot(data) |
positions |
Controls the positions of violin plots. | positions=[1, 2, 3] |
vert |
Controls orientation in older versions. | vert=False |
orientation |
Controls plot orientation. | orientation="horizontal" |
showmeans |
Displays the mean indicator. | showmeans=True |
showmedians |
Displays the median indicator. | showmedians=True |
showextrema |
Displays minimum and maximum indicators. | showextrema=True |
showquartiles |
Displays quartile indicators. | showquartiles=True |
widths |
Controls maximum violin width. | widths=0.5 |
bw_method |
Controls density estimation bandwidth. | bw_method=0.5 |
points |
Controls the number of points used to evaluate density. | points=100 |
quantiles |
Displays specified quantile positions. | quantiles=[[0.25, 0.5, 0.75]] |
20. Practical Example: Employee Salary Distribution
Let's create violin plots to compare the salary distribution of employees from three departments. We will display mean and median indicators and customize the violin colours.
Complete Python Program
import matplotlib.pyplot as plt
import numpy as np
np.random.seed(10)
it_department = np.random.normal(
loc=60000,
scale=10000,
size=100
)
hr_department = np.random.normal(
loc=45000,
scale=7000,
size=100
)
sales_department = np.random.normal(
loc=50000,
scale=12000,
size=100
)
data = [
it_department,
hr_department,
sales_department
]
fig, ax = plt.subplots(figsize=(9, 6))
result = ax.violinplot(
data,
showmeans=True,
showmedians=True,
showextrema=True
)
colors = [
"#93c5fd",
"#86efac",
"#fcd34d"
]
for body, color in zip(result["bodies"], colors):
body.set_facecolor(color)
body.set_edgecolor("#334155")
body.set_alpha(0.85)
ax.set_xticks(
[1, 2, 3],
["IT", "HR", "Sales"]
)
ax.set_title("Employee Salary Distribution")
ax.set_xlabel("Departments")
ax.set_ylabel("Salary")
ax.grid(axis="y", linestyle="--", alpha=0.4)
plt.show()
Expected Output
Employee Salary Distribution
IT Department : Blue Violin
HR Department : Green Violin
Sales Department : Yellow Violin
Each violin displays the estimated salary distribution,
mean indicator, median indicator and extrema.
The chart helps compare salary concentration
and variation across departments.
Understanding the Result
The IT department has a higher average salary in this example. The HR and Sales departments show different distributions and spreads. The violin shapes help understand where employee salaries are concentrated.
21. Best Practices for Violin Plots
- Use violin plots when distribution shape is important.
- Compare datasets using consistent scales.
- Use clear category labels.
- Display median and mean when useful.
- Choose a suitable bandwidth for density estimation.
- Avoid too many violin plots in one chart.
- Use colours that are easy to distinguish.
- Explain that violin width represents estimated density.
22. Common Mistakes and Solutions
| Common Mistake | Solution |
|---|---|
| Passing incorrectly structured data. | Pass numerical observations as a dataset or sequence of datasets. |
| Violin colours are not visible. | Customize the returned bodies collection. |
| Mean indicator is missing. | Set showmeans=True. |
| Median indicator is missing. | Set showmedians=True. |
| Density shape looks too smooth. | Review the bandwidth setting. |
| Category names are missing. | Use plt.xticks() or Axes tick methods. |
| Orientation parameter is not recognized. | Check the installed Matplotlib version. |
23. Practice Questions
Beginner Level
- What is a violin plot?
- Which Matplotlib function creates violin plots?
- What does the width of a violin represent?
- What is probability density?
- What is kernel density estimation?
Intermediate Level
- Explain the difference between a box plot and violin plot.
- What is the purpose of bw_method?
- How can you display the mean and median?
- How do you customize violin colours?
- What is the purpose of showextrema?
- How can you create a horizontal violin plot?
Practical Tasks
- Create a violin plot for student examination marks.
- Compare three different classes using violin plots.
- Display mean and median indicators.
- Customize violin colours and transparency.
- Create a horizontal violin plot.
- Compare employee salary distributions across departments.
Summary
A violin plot is a statistical chart used to visualize numerical data distribution. It combines information about probability density with useful statistical indicators. Wider areas represent higher estimated data density, while narrower areas represent lower estimated density.
Matplotlib provides the plt.violinplot() function
to create violin plots. It supports multiple datasets,
horizontal orientation, mean and median indicators,
extrema and quartile information.
The appearance of violin plots can be customized using the returned bodies collection. Parameters such as widths, bw_method and points help control the appearance and density estimation process.
Violin plots are useful when you need to understand the shape of a distribution and compare different datasets. They provide more information about density shape than a standard box plot.