I currently decompose time series with MSTL from statsmodels: https://www.statsmodels.org/dev/generated/statsmodels.tsa.seasonal.MSTL.html.
As far as I understood, MSTL assumes that the time series can be expressed as an additive decomposition. For reasons of interpretation, I would prefer a multiplicative model. Is it appropriate to take logs, do the decomposition, transform them back by exponentiation and interpret them as multiplicative?
