search

robdmc / pandashells

:panda_face: Bringing the python data stack to the shell prompt
Other
787 stars 26 forks source link

Change cdf to ecdf and use statsmodels #19

Closed robdmc closed 9 years ago

robdmc commented 9 years ago

Something like this should work

from statsmodels.distributions.empirical_distribution import ECDF
ecdf = ECDF(df_error.z)
x = sorted(df_error.z)
y = ecdf(x)
pl.plot(x, y)
pl.plot(x, 1-y)