YouTube Video Review: Time series analysis / forecasting by Dr A Sarkar, PhD

I had been somewhat reluctant to watch Dr Sarkar’s video in time series analysis / forecasting because it was based around the R programming language. I thought about it and decided it could not be that difficult to convert the code in R to Python, so I would give it a go. What Ii was to find out, however, is there are some functions in R that are not present in Python. For instance, the forecasting function Random Walk with Drift does not have an easy to translate substitute in Python. I did find an algorithm for Random Walk that I was able to modify for use in Python, but I still have not been able to find a Python equivalent to Random Walk. Nevertheless, I carried on with the video, endeavouring to learn as much as I could from this very knowledgeable man.

Types of statistics:-

1. Descriptive describes data.
2. Predictive answers the question of what happens in the future.
3. Prescriptive tells what actions to take based upon the observed data.

The components of time series:-

1. Trend covers the long term
2. Seasonality covers short term variations
3. Irregularity covers random movements