About the course

ITESO University Elendil TA GitHub repository Report an issue

What this site is

This is the public course hub for Time Series Forecasting.

It is intentionally built as a working reference:

  • lecture notes you can revisit during the semester
  • code-first examples you can reuse in projects
  • exercises that build muscle memory for real forecasting workflows

This course is taught at ITESO University as part of the undergraduate and graduate curriculum in Data Science, Financial Engineering, and related programs.

Note

If you only follow one rule: always compare against a baseline.
Fancy models without a baseline are just expensive opinions.

Who it’s for

You’ll do best in this course if you can:

  • manipulate data frames confidently (filter, mutate, group_by, join)
  • understand basic probability/statistics and linear regression
  • stay organized with reproducible work (projects, scripts, reports)
Important

This course is not “click-and-run forecasting”.
You will be expected to justify assumptions, diagnose models, and communicate tradeoffs.

How to use the site

Use the navbar as your map:

  • Modules: the core storyline (what you must learn)
  • Exercises: practice and feedback loops (what makes it stick)
  • More: optional refreshers (R tooling, stats, side quests)
  • About: this page

A practical routine:

  1. Read the module page.
  2. Run the code (don’t just skim it).
  3. Do the exercise.
  4. Write a short explanation of what changed and why.

Tools and conventions

We work primarily in R, using the tidy ecosystem for forecasting:

  • native pipe |> (read it as “then”)
  • tidy data + functional pipelines
  • forecasting as a workflow: split → fit → evaluate → iterate → communicate

Mathematics is included when it adds clarity, not as decoration.

If you come from Python, you can still follow along.
The mental model matters more than the syntax.

Elendil TA (AI code assistant)

Elendil TA is available from the navbar. Use it to:

  • debug errors and understand messages
  • translate ideas into R code
  • check your reasoning and interpretation

What not to do:

  • paste entire assignments and ask for the final answer
  • use it as a substitute for understanding your own results
Warning

If you can’t explain your model, you don’t own it.
Treat AI output as draft code that must be validated.

Feedback, corrections, and reuse

This site is maintained like a small open-source project.

  • Found a typo? Open an issue.
  • Want to improve an explanation? Propose a pull request.
  • Want to reuse material? Please cite the repository and keep attribution.

Credits and inspiration

This course is strongly inspired by Forecasting: Principles and Practice (FPP3) and the broader forecasting community.

Back to top