对曲线分布下的阴影区域着色 plot 不同 colors

Question

I'm using seaborn's kdeplot to draw the distribution of my data.

sns.kdeplot(data['numbers'], shade=True)

I want to divide the shaded area under the line into three parts, showing the "high" percentile and the "low" percentile. It would be ideal if I can color the shaded area with three different colors.

Any idea how I can go about doing that?

I want it to look something like the below where I can decide the cutoff value between the colors.

Answer 1

So I figured out how to do it. I would retrieve and x and y arrays from the seaborn plot, then use fill_between to color under the curve.

points = sns.kdeplot(data['numbers'], shade=True).get_lines()[0].get_data()

x = points[0]
y = points[1]

plt.fill_between(x,y, where = x >=0.75, color='r')
plt.fill_between(x,y, where = x <=0.1, color='g')
plt.fill_between(x,y, where = (x<=0.75) & (x>=0.1), color='y')

Answer 2

The only difference since the original answer was posted is that shade=True needs to be changed to shaded=False .

I also just wanted to share a slightly more complete answer for those who want to copy-paste.

import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

mean = 0
std = 1

np.random.seed(0)
x = np.random.normal(loc=mean, scale=std, size=10000)

points = sns.kdeplot(x, shade=False, bw_method=1.0, color='black').get_lines()[0].get_data()

x = points[0]
y = points[1]

plt.fill_between(x, y, color='y')
plt.fill_between(x, y, where = (x<=0.5*std) & (x >= -0.5 * std), color='g') # 
plt.fill_between(x, y, where = (x>=2*std) | (x<=-2*std), color='r')