Tutorials, models and utility functions for time series analysis.
Useful utility functions are explained in utility/learn_utilities.ipynb notebook. Also, functions we would use to handle time series are organized in utilities/utilities.py.
Notebook natural_language_processing/ds_beach_demand_nlp.ipynb shows how you can use NLP techniques to build a deep learning forecasting model. The model forecasts the demand for each (store, item) tuple using historical sales as well as item descriptions. In other words, the relationship between items is captured through item descriptions. The notebook is based on the notebook and youtube by Dr. Jeff Heaton. You can find his video on Youtube. I have modified the code, used my tested time series utility functions.