Search

📂

kdeplot

Status
Not started
Text

sns.kdeplot()

Text 1

Kernel Density Estimate plot for smooth distribution curves.

sns.kdeplot() is used to visualize the probability density function (PDF) of a continuous variable using Kernel Density Estimation (KDE). It provides a smooth alternative to histograms for understanding distribution shapes.

1. General Syntax

python
CopyEdit
sns.kdeplot(
    data=None,
    x=None,
    y=None,
    hue=None,
    weights=None,
    bw_method=None,
    bw_adjust=1,
    fill=False,
    common_norm=True,
    common_grid=False,
    cumulative=False,
    cut=3,
    clip=None,
    gridsize=200,
    legend=True,
    palette=None,
    alpha=None,
    linewidth=None,
    linestyle=None,
    thresh=0.05
)

2. Key Parameters

Parameter
Description
x, y
Input data for univariate or bivariate KDE.
hue
Grouping variable that will produce separate curves.
bw_adjust
Adjusts the bandwidth (smoothing); lower = more detail.
fill
Fills the area under the KDE curve.
cumulative
If True, shows the cumulative distribution.
common_norm
If False, normalizes each subgroup independently.
cut
Extends the curve beyond extreme data points (default is 3).
clip
Tuple to limit the KDE support range.
gridsize
Number of points used to evaluate the KDE curve.

3. Dataset Setup

We’ll use the penguins dataset from Seaborn.

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

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

Dataset Columns

  • species: Penguin species.
  • bill_length_mm: Length of the bill.
  • bill_depth_mm: Depth of the bill.
  • flipper_length_mm: Flipper length.
  • body_mass_g: Body mass.
  • sex: Gender.

4. Basic KDE Plot

Plot distribution of bill length


sns.kdeplot(data=data, x="bill_length_mm")
plt.title("KDE Plot of Bill Length")
plt.show()

✅ Explanation

  • Shows a smooth distribution curve for bill_length_mm.

5. KDE Plot with Shaded Area (fill=True)

python
CopyEdit
sns.kdeplot(data=data, x="bill_length_mm", fill=True)
plt.title("Filled KDE Plot of Bill Length")
plt.show()

✅ Explanation

  • The area under the curve is shaded, making the shape more visible.

6. Grouping by Category (hue)

python
CopyEdit
sns.kdeplot(data=data, x="bill_length_mm", hue="species", fill=True)
plt.title("KDE Plot of Bill Length by Species")
plt.show()

✅ Explanation

  • Shows separate KDE curves for each penguin species.

7. Adjusting Smoothing (bw_adjust)

python
CopyEdit
sns.kdeplot(data=data, x="bill_length_mm", bw_adjust=0.5, fill=True)
plt.title("KDE Plot with Lower Bandwidth (More Detail)")
plt.show()


python
CopyEdit
sns.kdeplot(data=data, x="bill_length_mm", bw_adjust=2, fill=True)
plt.title("KDE Plot with Higher Bandwidth (More Smoothing)")
plt.show()

✅ Explanation

  • Lower bw_adjust = more peaks and valleys (may overfit).
  • Higher bw_adjust = smoother curve (may oversmooth).

8. Cumulative KDE Plot

python
CopyEdit
sns.kdeplot(data=data, x="bill_length_mm", cumulative=True)
plt.title("Cumulative KDE Plot of Bill Length")
plt.show()

✅ Explanation

  • Shows the cumulative distribution (like a CDF).

9. Clipping the Curve

python
CopyEdit
sns.kdeplot(data=data, x="bill_length_mm", clip=(35, 50))
plt.title("KDE Plot with Clipping Range (35–50 mm)")
plt.show()

✅ Explanation

  • Restricts the x-axis range of the KDE.

10. Horizontal KDE

python
CopyEdit
sns.kdeplot(data=data, y="bill_length_mm", fill=True)
plt.title("Horizontal KDE Plot")
plt.show()

✅ Explanation

  • Same KDE curve, but along the y-axis.

11. Customizing Appearance

python
CopyEdit
sns.kdeplot(
    data=data,
    x="bill_length_mm",
    hue="species",
    fill=True,
    alpha=0.4,
    linewidth=2,
    linestyle="--"
)
plt.title("Customized KDE Plot")
plt.show()

✅ Explanation

  • Custom alpha for transparency, linewidth, and linestyle.

12. 2D KDE Plot (Bivariate)

python
CopyEdit
sns.kdeplot(data=data, x="bill_length_mm", y="bill_depth_mm", fill=True, cmap="Blues")
plt.title("2D KDE Plot of Bill Length and Depth")
plt.show()

✅ Explanation

  • Shows density contours for joint distribution of two variables.

13. 2D KDE with Hue

python
CopyEdit
sns.kdeplot(
    data=data,
    x="bill_length_mm",
    y="bill_depth_mm",
    hue="species",
    fill=True
)
plt.title("2D KDE by Species")
plt.show()

✅ Explanation

  • One KDE surface per species, filled with different colors.

14. Final Example: Fully Customized KDE

python
CopyEdit
plt.figure(figsize=(10, 6))
sns.kdeplot(
    data=data,
    x="bill_length_mm",
    hue="species",
    fill=True,
    bw_adjust=1.2,
    alpha=0.5,
    linewidth=2,
    palette="Set2"
)
plt.title("Smooth KDE Plot of Bill Length by Species")
plt.grid(axis="y", linestyle="--", alpha=0.6)
plt.show()

✅ Features in This Plot:

  • Custom bandwidth
  • Shaded curves
  • Hue separation by species
  • Larger plot size
  • Light grid for clarity

Conclusion

✅ sns.kdeplot() is excellent for:

  • Exploring distribution shapes
  • Visualizing group-wise densities
  • Creating 2D smoothed plots

✅ It's a smooth and flexible alternative to histograms and is especially useful for probabilistic insights in exploratory data analysis.

Master KDE plots to elevate your distribution analysis in a visually elegant way! 🚀