4. Distribution Plot

Let’s use a different plotting style, bmh.

%matplotlib inline
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
import pandas as pd

plt.style.use('bmh')

np.random.seed(37)

4.1. Basic

Here are two basic density plots using different bandwith methods.

sigma = 1.0
mu = 100.8

x = sigma * np.random.randn(1000) + mu
s = pd.Series(x)

fig, ax = plt.subplots(1, 2, figsize=(15, 3))

_ = s.plot(kind='kde', bw_method='scott', ax=ax[0])
_ = s.plot(kind='kde', bw_method='silverman', ax=ax[1])

_ = ax[0].set_title('Density plot, bandwith=scott')
_ = ax[1].set_title('Density plot, bandwith=silverman')

4.2. Multiple densities in a single plot

Multiple densities in a single plot may be achieved through Pandas dataframes.

df = pd.DataFrame({
    'x': 1.0 * np.random.randn(1000) + 100,
    'y': 2.0 * np.random.randn(1000) + 105
})

fig, ax = plt.subplots(figsize=(10, 5))

_ = df.plot(kind='kde', ax=ax)
_ = ax.set_title('Multiple densities')

4.3. Density with cumulative distribution plot

This plot overlays density and cumulative plots using Seaborn.

sigma = 1.0
mu = 100.8

x = sigma * np.random.randn(1000) + mu

fig, ax1 = plt.subplots(figsize=(10, 5))

_ = sns.kdeplot(x, ax=ax1, color='r', label='density')

ax2 = ax1.twinx()

_ = sns.kdeplot(x, ax=ax2, cumulative=True, color='g', label='cumulative')

_ = ax1.set_ylabel('probability', color='r')
_ = ax2.set_ylabel('cumulative', color='g')

_ = ax1.tick_params(axis='y', labelcolor='r')
_ = ax2.tick_params(axis='y', labelcolor='g')

_ = ax1.legend().set_visible(False)
_ = ax2.legend().set_visible(False)

_ = ax1.set_title('Density with cumulative distribution plots')

4.4. Histogram overlay

sigma = 1.0
mu = 100.8

x = sigma * np.random.randn(1000) + mu
s = pd.Series(x)

fig, ax = plt.subplots(figsize=(15, 3))

ax1, ax2 = ax, ax.twinx()

_ = s.plot(kind='hist', bins=15, alpha=0.5, color='r', ax=ax1)
_ = s.plot(kind='kde', bw_method='scott', color='b', ax=ax2)


_ = ax.set_title('Density plot with histogram')