The idea behind this project was to better understand how to build time series models in python and try it on a real data. What is the process for selecting best model in time series, which criterias to use, etc.
I used seasonal and non seasonal ARIMA models, calculated model accuracy in and out of sample, utilized cross validation technique, compared model performance predicting a series of observations vs one observation at a time, plotted forecasted values vs actuals to understand how ARIMA works.