This repo contains the notebook used to generate the figures in this article on MSTL.
In the notebook I show how to decompose a time series with multiple seasonal components using an algorithm called Multiple Seasonal-Trend decomposition using Loess (MSTL) in Python. I demo an implementation of MSTL that I contributed to Statsmodels and apply it to an electricity demand time series.
Create a virtual environment and pip install the requirements.
pip install -r requirements.txt