Summary statistics

- split the data in more partitions
- compare the mean and the variance of each group
- if the the difference is statistically not significant, then the serie is likely non-stationary


Also use ANOVA test
These calculations can be done from scipy and statistical module stats (function: f_onway).

Statistical test

- Augmented Dickey-Fuller (ADF)
from statsmodels.tsa.stattools with the function adfuller

- Kwiatkowski-Phillips-Schmidt-Shin
from statsmodels.tsa.stattools with the function kpss