getting only the last plot - matplotlib

Question

when plotting by below code I am getting c,d,e plots but I am getting only the last plot for plt.plot
def normalize(x):
return (x - x.min(0)) / x.ptp(0)
c=sns.distplot(mk[0]['mass'], hist=True, label='p', rug=True)
d=sns.distplot(mk[1]['mass'], hist=True, label='q', rug=True)
e=sns.distplot(mk[2]['mass'], hist=True, label='r', rug=True)
datadist=[c,d,e]
xd=dict()
yd2=dict()
for i in datadist:
line = i.get_lines()[0]
xd[i] = line.get_xdata()
yd = line.get_ydata()
yd2[i] = normalize(yd)
plt.plot(xd[c], yd2[c],color='black')
plt.plot(xd[d], yd2[d],color='yellow')
plt.plot(xd[e], yd2[e],color='green')

Answer 1

sns.distplot() returns the ax (the subplot) on which the histogram was drawn. All 3 are drawn on the same subplot, so the return value is the same three times.
The array lines = ax1.get_lines() contains exactly 3 elements: one for each of the kde curves, so you can extract them as follows:
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
def normalize(x):
return (x - x.min(0)) / x.ptp(0)
fig, (ax1, ax2) = plt.subplots(ncols=2, figsize=(14, 4))
sns.distplot(np.random.randn(30) + 10, hist=True, label='p', rug=True, ax=ax1, color='black')
sns.distplot(np.random.randn(30) + 15, hist=True, label='q', rug=True, ax=ax1, color='gold')
sns.distplot(np.random.randn(30) + 20, hist=True, label='r', rug=True, ax=ax1, color='green')
for line in ax1.get_lines():
ax2.plot(line.get_xdata(), normalize(line.get_ydata()), color=line.get_color())
plt.show()
Now, if you just want the kde-curves and "normalize" them, you could use scipy.stats import gaussian_kde:
import numpy as np
import matplotlib.pyplot as plt
from scipy.stats import gaussian_kde
def normalize(x):
return (x - x.min(0)) / x.ptp(0)
fig, ax = plt.subplots(figsize=(12, 4))
mk0mass = np.random.randn(30) + 10
mk1mass = np.random.randn(30) + 15
mk2mass = np.random.randn(30) + 20
all_mkmass = [mk0mass, mk1mass, mk2mass]
x = np.linspace(min([mki.min() for mki in all_mkmass]) - 2,
max([mki.max() for mki in all_mkmass]) + 2, 1000)
for mki, color in zip(all_mkmass, ['black', 'gold', 'green']):
kde = gaussian_kde(mki)
yd = normalize(kde(x))
ax.plot(x, yd, color=color)
ax.fill_between(x, 0, yd, color=color, alpha=0.3)
plt.show()