Question

why hatches are at the wrong place when I change y axis to Log scale?

if I remove log scale they are inside histogram-columns , and outside for log scale

from scipy.stats import norm
import matplotlib.mlab as mlab
import numpy as np
import matplotlib.pyplot as plt



data=np.genfromtxt(FileName)

Flow=data[:,1]

Max=np.amax(Flow)
Min=np.amin(Flow)


n, bins=np.histogram(Flow, 15)


n, bins, patches = plt.hist(Flow, 15, color= 'b',log=True, histtype='step',hatch='/')


(mu, sigma) = norm.fit(Flow)

y = mlab.normpdf( bins, mu, sigma)

plt.xlabel('Flow (m3/s)')

plt.ylabel('Number of Observations')

plt.title(r'$\mathrm{Flow:}\ \mu=%.3f,\ \sigma=%.3f, Min=%.3f, Max=%.3f$' %(mu, sigma, Min, Max))



plt.show()

Answer 1

Because when draw the step lines, the first element of Y axis data is always 0, and log scale cannot display 0, so the first point of the step lines is dropped.

You can try to add a very small positive value to the data:

n, bins, patches = plt.hist(Flow+1e-12, 15, color= 'b',log=True, histtype='step',hatch='/')