Seaborn sns.barplot() – Categorical Aggregation with Bars

sns.barplot() is used to visualize the mean (or other aggregate statistics) of a numerical variable for different categorical variables. It is useful for comparisons across categories.

1. General Syntax

python
CopyEdit
sns.barplot(
    data=None,
    x=None,
    y=None,
    hue=None,
    order=None,
    hue_order=None,
    estimator=None,
    ci=95,
    n_boot=1000,
    units=None,
    seed=None,
    orient=None,
    color=None,
    palette=None,
    saturation=0.75,
    width=0.8,
    errcolor=".26",
    errwidth=None,
    capsize=None,
    dodge=True
)

2. Key Parameters

3. Dataset Setup

We will use the titanic dataset from Seaborn.

python
CopyEdit
import seaborn as sns
import matplotlib.pyplot as plt

# Load dataset
data = sns.load_dataset("titanic")
data.head()

Dataset Columns

• survived: Whether the passenger survived (0 = No, 1 = Yes).
• pclass: Passenger class (1st, 2nd, 3rd).
• sex: Gender.
• age: Age in years.
• fare: Ticket fare.
• embark_town: Port of embarkation.

4. Basic Bar Plot

Plot the average fare for each passenger class (pclass)

python
CopyEdit
sns.barplot(data=data, x="pclass", y="fare")
plt.title("Average Fare by Passenger Class")
plt.show()

Explanation

• The height of the bars represents the average fare.
• Default 95% confidence interval (CI) is shown as error bars.

5. Removing Confidence Intervals (ci=None)

python
CopyEdit
sns.barplot(data=data, x="pclass", y="fare", ci=None)
plt.title("Bar Plot Without Confidence Intervals")
plt.show()

Explanation

• Removing confidence intervals makes the plot cleaner.

6. Changing the Aggregation Method (estimator)

Use median instead of mean

python
CopyEdit
import numpy as np

sns.barplot(data=data, x="pclass", y="fare", estimator=np.median)
plt.title("Median Fare by Passenger Class")
plt.show()

Explanation

• By default, sns.barplot() calculates mean values.
• estimator=np.median changes the calculation to median.

7. Grouping Data with hue

Compare fares across sex within each pclass

python
CopyEdit
sns.barplot(data=data, x="pclass", y="fare", hue="sex")
plt.title("Average Fare by Passenger Class and Gender")
plt.show()

Explanation

• Uses different colors for male and female groups.

8. Adjusting Bar Width (width)

python
CopyEdit
sns.barplot(data=data, x="pclass", y="fare", width=0.5)
plt.title("Bar Plot with Narrower Bars")
plt.show()

Explanation

• width=0.5 reduces bar width.

9. Changing Color Palettes (palette)

Use coolwarm palette

python
CopyEdit
sns.barplot(data=data, x="pclass", y="fare", hue="sex", palette="coolwarm")
plt.title("Custom Color Palette")
plt.show()

Explanation

• Uses Seaborn’s predefined color palettes.

10. Adding Error Bar Caps (capsize)

python
CopyEdit
sns.barplot(data=data, x="pclass", y="fare", capsize=0.2)
plt.title("Bar Plot with Error Bar Caps")
plt.show()

Explanation

• Caps make error bars easier to interpret.

11. Customizing Error Bars (errwidth and errcolor)

python
CopyEdit
sns.barplot(data=data, x="pclass", y="fare", errwidth=2, errcolor="red")
plt.title("Customized Error Bars")
plt.show()

Explanation

• Increases error bar width and changes color to red.

12. Using dodge=False to Stack Bars

python
CopyEdit
sns.barplot(data=data, x="pclass", y="fare", hue="sex", dodge=False)
plt.title("Stacked Bars Instead of Side-by-Side")
plt.show()

Explanation

• Prevents bars from being separated when using hue.

13. Horizontal Bar Plot

python
CopyEdit
sns.barplot(data=data, y="pclass", x="fare")
plt.title("Horizontal Bar Plot of Fare by Passenger Class")
plt.show()

Explanation

• Swaps x and y to create a horizontal bar plot.

14. Rotating X-axis Labels

python
CopyEdit
sns.barplot(data=data, x="embark_town", y="fare")
plt.xticks(rotation=45)
plt.title("Rotated X-axis Labels")
plt.show()

Explanation

• Prevents overlapping labels on the x-axis.

15. Adjusting Figure Size

python
CopyEdit
plt.figure(figsize=(10, 6))
sns.barplot(data=data, x="pclass", y="fare")
plt.title("Larger Figure Size")
plt.show()

Explanation

• Improves readability for large datasets.

16. Final Example: Fully Customized Bar Plot

Features in This Plot

✔ Hue by sex

✔ Caps on error bars

✔ Custom width and error bar thickness

✔ Larger figure size

✔ Grid lines for clarity

Conclusion

✅ sns.barplot() is a great tool for visualizing category-based aggregation.

✅ It provides clear comparisons using means, medians, or other statistical measures.

✅ Highly customizable, supporting hue-based grouping, confidence intervals, color palettes, error bars, and grid styling.

Mastering bar plots will enhance your ability to present and compare numerical data effectively! 🚀