Frequência relativa do histograma

# assuming that mydata is an numpy array
 ax.hist(mydata, weights=np.zeros_like(mydata) + 1. / mydata.size)
Ananda Gorck Streit