Energy System Modeling with R • energyRt

A post-conference online series following useR! 2026 — live Zoom sessions on building, solving, and exploring energy-system models in R.

The course focuses primarily on energyRt. It also introduces the broader optimal2050 ecosystem of R packages from the perspective of the “LEGO-lization” of energy modeling: assembling transparent and reproducible modeling workflows from interchangeable, reusable components.

With energyRt, an energy system is described as a collection of R objects. The package compiles and solves the resulting optimization model using a mathematical-programming backend, and returns the results to R as tidy tables for analysis and visualization.

Register: sign up on Zoom » — four Fridays, 7–28 August 2026, 12:00 UTC.

An open work in progress. Most of the optimal2050 packages are still in active development and not yet formally released. This course shares the ideas behind an ongoing project as much as it teaches finished software—expect a few rough edges, and bring your questions and critique. We are keen to discuss the approach, improve it together, and welcome new collaborators.

How the course works. The four meetings alternate between teaching and follow-up. Sessions 1 and 3 introduce new material with live demonstrations; Sessions 2 and 4 are follow-up sessions where we review your exercises, troubleshoot code, and answer questions. Between meetings you work through short exercises at your own pace. Materials and exercises are published on the course site as the series progresses, so the sessions build on one another and you can catch up on anything you miss.

Schedule

All sessions run at 12:00 UTC (New York 08:00 · Berlin 14:00 · New Delhi 17:30 · Beijing 20:00) and last approximately 1.5 hours.

Session 1 — LEGO-lization of energy modeling and energyRt basics

Friday, 7 August 2026

A brief overview of energy-system models and their place in the broader energy-economy-environment modeling landscape
An overview of the optimal2050 ecosystem and modular, “LEGO-like” modeling workflows
An end-to-end workflow demonstration—from defining a model to solving a tiny model live and exploring the results
The basic concepts and structure of energyRt
Setting up and solving a simple energy-system model in R
Introduction to the hands-on exercises

Session 2 — Exercise follow-up

Friday, 14 August 2026

Review and discussion of the exercises from Session 1
Building an example model “brick by brick”
Comparison of solutions and modeling approaches
Policy-scenario examples, such as emissions constraints and renewable-energy targets
Questions arising from participants’ practical work

Session 3 — Advanced energyRt applications

Friday, 21 August 2026

Advanced energy-system modeling examples
Reproducing a large-scale, real-world open-source energy-system model in energyRt (the specific model will be announced before the session)
Model structure, data organization, and scenario design
Practical considerations for larger, multi-region applications
Using alternative mathematical-programming backends (GLPK, HiGHS/Julia, Pyomo, GAMS, and license-free remote solving via NEOS)

Session 4 — Questions and troubleshooting

Friday, 28 August 2026

Open questions about energyRt and the course materials
Troubleshooting participants’ models and code
Discussion of modeling challenges and possible solutions
Guidance on extending the examples to participants’ own applications
Suggested next steps and further resources

Who it is for

The series is intended for researchers, analysts, students, and practitioners interested in energy-system modeling with R. Familiarity with R will be helpful, but prior experience with energyRt is not required.

Preparation

Nothing is required for the first session—no installation and no setup. Session 1 is a live demonstration; you can simply join and watch.

The hands-on exercises are done on your own time between the sessions. To take part in those, you will need R (version 4.3 or newer), energyRt, and at least one supported solver (RStudio recommended)—and you are equally welcome to follow the whole series without installing anything.

Everything is published as you go, so you can work at your own pace:

R: setup and resources — where to get R, RStudio, and Quarto; short introductions to the language; the packages worth installing first (dplyr, ggplot2, sf); and the cheatsheets to keep to hand.
Installing energyRt — the package itself plus at least one solver backend.
Your first energy model — the first exercise: start from the smallest model that can exist and grow it one object at a time until it burns coal and emits CO₂.
Building blocks of an energyRt model — a reference deck on the new*() constructors and the system around them, used in the sessions and written to be browsable afterwards.

New to R? The first of those pages is the place to start, and the sessions assume no prior experience with energyRt.

All of it lives in the course materials site (source on GitHub).

Questions and discussion

Questions between sessions are welcome, and best asked in the open where the answer helps everyone:

Ask here: the course Q&A thread on Mastodon

Reply with anything—an install that will not cooperate, a modeling question, a result you cannot explain, or something you have built and want to show. Beginner questions are explicitly welcome.

If you do not have a Mastodon account, registering takes about a minute: mastodon.social is the largest general server and a fine default. Mastodon is decentralized, so an account on any server can reply to the thread. Once you are in, search for the full handle @optimal2050@mstdn.science and follow @optimal2050 for course updates; tag your own posts with #optimal2050.

Registration

Participants are encouraged to bring their questions, code, and modeling problems to the final session. We look forward to seeing you online.