Search code examples
pythontime-seriesstatsmodelstrend

Code seasonal_decompose from statsmodels with only native python

I need to get the same result of the following code:

from statsmodels.tsa.seasonal import seasonal_decompose

# the values are just a list of numbers ([589, 561, 640, 656, ...])
values =  list(df['Monthly milk production'].as_matrix().transpose())

trend = seasonal_decompose(values, freq = 12).trend

but without using the statsmodels module

I wonder if there is just a formula that computes the trend values so I can replace that formula with native python.

Solution

  • You can check the actual code of statsmodels.tsa.seasonal.

    It is online at https://github.com/statsmodels/statsmodels/blob/master/statsmodels/tsa/seasonal.py

    All in Python