How to obtain seasonally-adjusted time series data using STL in Python

Question

On the section "STL decomposition" in the 2nd edition of Forecasting: Principles and Practice, it says that the seasadj() function can be used to compute the seasonally adjusted series but it does not say how this seasonally adjusted series is computed. I'm wondering how to do this in python's statsmodel package as there is no such seasadj function.

If I have an STL decomposition like so:

would the seasonally adjusted series be the trend + remainder? Just the trend?

In python:

from statsmodels.tsa.seasonal import STL

stl_decomp = STL(series, period=12, seasonal=7).fit()
stl_seas_adj = stl_decomp.trend + stl_decomp.resid

Is this the correct way to compute the seasonally adjusted data?

Answer 1

From the statsmodels docs, stl.fit() returns a DecomposeResult object. This has the estimated seasonal component of the time series decomposition. Subtracting this from the original time series should then provide you with a seasonally-adjusted time series.