You are supposed to work with daily data on Russian spot elecricity...

You are supposed to work with daily data on Russian spot elecricity market. Choose one of the four time series from the dataset you are interested to work with.

1. (15 points) Visualize the series itself, the series of ordinary and seasonal differences, the components of the series, the ordinary and partial autocorrelation functions. Comment on the created plots.
2. (15 points) Prepare the data: fill in the missing values, remove outliers and correct for structural breaks if needed. Comment on the choice of the methods and the parameters.
3. (5 points) Is the series stationary? Support your answer with a suitable test.
4. (5 points) If it is reasonable to apply any transformation to the original series? Motivate your choice.
5. (50 points) Divide the series into test and train sets. Evaluate a set of models/algorithms on a test set. As a minimum, you should consider: SARIMA, ETS, OLS + a comparator of your choice. You are welcome to consider additional models as well. Provide a comment on the choice of model parameters!
6. (5 points) Choose the best model.
7. (5 points) Plot forecasts two years ahead by reevaluating the best model over the full dataset.
Comments
The work should be submitted as a report in pdf or html format. The work there should consist of a text with graphs and tables needed to support your reasoning, the code should be in the Appendix. The report (without the Appendix) should be no longer than 10 pages.
This work implies some creative approach. In each case, you can do more than the specified minimum. Note, that you are welcome to use any techniques that are not mentioned in the course