How can a color fill be done between two histogram lines with matplotlib? - matplotlib

Question

I have two histograms (signal and background) and each of them have upper and lower uncertainty bounds. I would like to plot each histogram as using only a colored edge with no fill and then to plot the upper and lower uncertainty histograms with their own colored edges with no fill, and then to have the region between the upper and lower histograms filled, as a sort of undertainty band. How might this be achieved?
Here's the very rough code I have just now (which can be used in Jupyter):
import matplotlib.pyplot as plt
import numpy as np
from pylab import normal
import seaborn as sns
sns.set(context="paper", font="monospace")
import warnings
warnings.filterwarnings('ignore')
%matplotlib inline
plt.rcParams["figure.figsize"] = [8, 8]
plt.rc("text", usetex=True)
plt.rc("font", family="serif")
s = normal(1, .2, 5000)
s_up = normal(1, .2, 5500)
s_dn = normal(1, .2, 4600)
b = normal(2, .2, 2500)
ns, bins, patches = plt.hist(
[s, s_up, s_dn, b],
color=['r', 'g', 'g', 'b'],
label=['signal nominal', 'signal upper bound', 'signal lower bound', 'background'],
alpha=0.5,
linewidth=1,
histtype='stepfilled');
plt.setp(patches[0], edgecolor='r')
plt.setp(patches[1], edgecolor='b')
plt.legend();
plt.show();

Answer 1

Something like that?
s = normal(1, .2, 5000)
s_up = normal(1, .2, 5500)
s_dn = normal(1, .2, 4600)
bins = np.linspace(0., 2., 21)
n_up,b_up,p_up = plt.hist(s_up, bins=bins, bottom=0, linewidth=1, histtype='stepfilled', facecolor='none', edgecolor='red', linestyle='--', label='signal upper bound')
n_dn,b_dn,p_dn = plt.hist(s_dn, bins=bins, bottom=0, linewidth=1, histtype='stepfilled', facecolor='none', edgecolor='green', linestyle='--', label='signal lower bound')
n,b,p = plt.hist(s, bins=bins, bottom=0, linewidth=2, histtype='stepfilled', facecolor='none', edgecolor='k', label='signal nominal')
plt.bar(x=b_up[:-1], height=n_up-n_dn, bottom=n_dn, width=np.diff(b_up), align='edge', linewidth=0, color='red', alpha=0.25, zorder=-1, label='uncertainty band')
plt.legend()