- Modelling and forecasting of price and volatility in power markets
- Modelling and forecasting of price and volatility in oil and natural gas markets
- Analysing investment decisions in renewable generation
- Analysing exploratory- and extraction-drilling decisions for offshore oil and gas
- The effects of C02 taxes and cap-and-trade allowances on energy markets
- The effects of taxation rules on oil and gas extraction
- Week 5: Wednesday January 18th to Friday January 20th.
- Week 9: Wednesday February 22nd to Friday February 24th.
- Week 14: Wednesday March 29th to Friday March 31st.
- Work on assignments in small groups (2-3 people). (The term project is limited to groups of 2, so it might be useful to also work in groups of 2 for the assignments.) Consult and collaborate with other students in your
*working groups*(of 4-8 students) - Do a google search to see if you can find an easy solution.
- Stackoverflow is probably one of the best forums for getting help (also for other programming langauges)
- Bring questions to the in-person sessions.
- Send me an email.

# Energy Industry Analytics, ENE 434

## NHH

### Spring 2023

### Lecturer: Johannes Mauritzen

#### Introduction

The best way of getting hold of me is by email:

johannes[dot]mauritzen[at]ntnu[dot]no

I will try to get back to you within 1 working day (24 hours)

### About the course

This course is an introduction in using data to analyse, describe and help make decisions within the energy industry and energy markets. The course covers both the power industry and oil and gas extraction.

In the course, students will learn how to use data to answer questions of how, when and where investments in energy generation and extraction take place. Students will also learn how to use data to analyse and make decisions related to power- and petroleum-industry operations. A particular focus of the course will be using data to analyse changes in market structure induced by changing regulations or the shift towards low-carbon technologies. Students will use the statistical software R to organise, clean, visualise and model energy data.

### Topics:

### Organisation

##### The course is officially organised in 3 intensive 3-day sessions of 2x2-hour class periods

##### 1015-1200, 1415-16

##### Sessions will be interactive, and involve working through “labs” that continue into group assignments.

In practice, you can work through the labs when it suits you and at your own pace throughout the semester.

I have placed videos and audio commentary integrated into the labs. I encourage you to work through the labs in groups.

### If you have a question:

You most likely will have questions and problems. Here is how you get help in the preferred order:

### Official course info from NHH

### Course Plan (tentative, can change through the semester.)

Date | Theme | Literature | Lab |
---|---|---|---|

18.1 | Intro to R, Cleaning data, Oil and Gas | RDS | Pre-Lab Lab 1 |

19.1 | Split-apply-combine: Working with large data sets. Petroleum company financials | RDS | Lab 2 |

20.1 | Relationsal data. Petroleum exploration and company financials. | RDS | Lab 3 |

22.2 | A review of time series statistics, with applications to power markets | FPP3 | Lab 4 |

23.2 | Power Markets: Univariate modelling and forecasting of time series | FPP3 | Lab 5 |

24.2 | Dynamic regression and forecasting: Modelling multiple power market variables. | FPP3 | Lab 6 |

29.3 | Solar power, learning curves, and non-linear regression | ISL | Lab 7 |

30.3 | The determinants of solar power costs in California: Regularisation and semi-parametric models. | ISL | Lab 8 |

31.3 | Nonlinear models: A poisson model of petroleum drilling and decision analysis of a wind power investment. | Lab 9 |

- Garrett Grolemund and Hadley Wickham (2017). “R for Data Science (RDS)”
- James, Witten, Hastie and Tibshirani, 2nd edition (2021). “Introduction to Statistical Learning” (ISL)
- Rob J Hyndman and George Athanasopoulos (2020). “Forecasting: Principles and Practice” (FPP3)

### Literature

We will make use of three texts in this course, all available freely online (see below). In general, I will refer to relevant parts of each text in the labs. But as background, you can read Ch. 1 and 2 of R for Data Science, Ch. 1 of Forecasting: Principles and Practice, and Ch 1 and 2 of statistical learning, plus ch. 3 for a refresher on linear statistical models.

- Grading is based on an class project - you can work alone or with one other person.
- There are in addition a total of 9 assignments, to be found at the end of each of the 9 labs. You will be forming larger
*working groups*of between 4 to 8 people that I encourage you to collaborate with and get help from. - You should complete all 9 assignments.
- In order to get credit for the course you need to submit the 3 assignments named below - one common submission for each working group. (You are encouraged to work in a smaller group (2 people) or independently for each assignment, but the
*working group*is a forum to get help and feedback, and you will choose one common assignment to turn in for each working group. I will give feedback on this common assignment. - The assignments should be submitted through canvas.
- The 3 assignments you turn in should be in the form of a r-markdown notebook, rendered as a PDF.
- You must turn in assignment 3 (from lab 3) from the first module (deadline 20.2.2023, 1200)
- You must turn in assignment 6 from the second module (deadline 28.3.2023, 1200)
- You must turn-in assignment 8 and 9 from the third module (deadline 21.4.2023, 1200)
- The handed-in assignments are cumulative, so you will need to have worked through the previous labs and assignments in order to complete them.
- I will provide feed-back on each assignment that is turned in.
- If you experience problems rendering your file as a pdf, you can try rendering as a HTML file, and then printing to pdf from a webbrowser.
- Criteria for the term project:
- Deadline as posted on the NHH examination plan
- You can start the project any time you wish: In general, the earlier the better.
- You can work alone or with one other person (groups of 2)
- The project should be written with r-markdown and rendered as a PDF.
- There is no strict length requirements, but a project should aim to be between 2.000-6.000 words. Longer projects do not count negatively in the evaluation, but neither does it count positively.
- The assignment will be a mix of a typical term paper and a lab report
- You have wide latitude in deciding the subject and methods in the project, as long as the subject and data is directly related to the energy industry.
- I might suggest you explore some of the data sources that I have posted below if you are having a hard time landing on a subject, but you should not feel limited to these sources. (Not all of the links are up-to-date).
- You should have introduction and background sections that explain the goals of your project, background info on the markets and data used, a summary of findings, and a short review of literature (1-2 relevant academic articles as well as other relevant sources).
- You should include code (and explanation of code when necessary) for the cleaning of the data, visualisation, model set-up and execution, as well as presentation of results.
- You do not need sections explicitly for theory and methods, but you should discuss theory and methods.
- You should explain and describe your results.
- You should provide a concluding section where you summarise your results, discuss their implications, weaknesses of the study, etc.
- Students should make use of at least one statistical model or technique beyond simple linear regression. Students can make use of the models covered in this course, but are not limited to those models. These models include time series ARIMA models, GARCH models, advanced seasonality models (fourier analysis), regularisation models (ridge regression, lasso), log-linear forecasting, and non- parametric and semi-parametric models (splines, loess, GAM).
- I will not provide students with suggested topics or--beyond the requirements listed above--a template or example term project. This is meant to encourage independence and creativity. In other words, the ability to find an interesting topic and make independent decisions on structure and substance are an important part of the term project.
- I encourage you to start working on your term projects as early as possible in the semester. Feel free to ask questions about the term project in the meet-up sessions. However,
**after May 1st, should be considered an exam period, and I will have to limit the help I give to general questions of formatting or structure.**

### Assignments, project and grading

### General comments on expectations and evaluations

The expectations for the project are different from a typical thesis or term paper. There are elements of a formal article that I would like to see — a well written introduction, some limited review of literature (find a handful of academic articles, reports, etc), a conclusion. But the project is intended to be loosely organised, where instead of just showing end results, the student also show the analysis that led to your results — exploratory graphs, summary statistics, statistical tests, and various hypothesis exploration. This project could be complimentary to writing a full article or thesis — serving as a well structured "lab notebook" detailing the research/analysis process.

Grading of such an open-ended project will necessarily be somewhat diffuse and somewhat subjective, and I rely to a high degree on the evaluation of a well-qualified external examiner. We strive to be fair and reasonable in the grading, but I can not ahead of time give a detailed break-down about what the exact break-points are for the different grades. Just to properly set expectations, I would refer to the general guidelines at NHH that a grade of A is typically reserved for truly exceptional projects. You can have a very good project without anything necessarily being "wrong", and still not get an A.

### Feedback

Below is an (unofficial) feedback form for the course. Responses are anonymous and I appreciate all constructive comments, suggestions and critiques.

# Resources

Here is a list of data sources I have run across over the years.

Here is a list of sources for learning more about statistics and machine learning in general.

Here is a list of sources for learning more about Bayesian analysis and statistics

Here is a list of sources for learning more about programming in r.

Here is a list of sources for learning more about programming in python.