Orvessa is an imaginary island whose farms grow candyflower. The candyflower is exported raw or processed into candy, and plantations are expanding into the island’s rainforest. The government publishes the kinds of data Trase works with: production and deforestation by municipality, a register of farms, silos and factories, distances between municipalities, and customs records for every export shipment.
The question the model answers is the same one at the heart of Trase:
For each export shipment, where was the commodity produced, and how much of that place’s deforestation should the shipment carry?
All of the data is synthetic and fits in five small files. That means you can read every number, follow every step and change anything, without needing accounts, cloud storage or databases.
What you can do with it
- Run a complete supply chain model in a few seconds with a single command.
- Follow one shipment from start to finish: see it matched to a storage hub, linked to the municipalities that supplied it, and given its share of deforestation.
- See how the main methods work in practice: a decision tree that uses what is known about each exporter, linear programming that links production areas to export hubs by shortest distance, and the attribution of deforestation to trade flows.
- Explore the results in a step-by-step notebook with charts, including a Sankey diagram of flows from municipality to destination country.
- Experiment: change the data or the model and see how the results respond. A set of graded exercises guides you, from tracing a single shipment to putting bounds on uncertain estimates.
Who it’s for
Researchers, students, analysts and anyone curious about how supply chain mapping works under the hood. It assumes a basic familiarity with supply chain concepts and some experience running Python. The documentation explains everything else.
What’s included
- The model code, with comments throughout
- A simplified version of the engine Trase uses to build supply chain models (SEI-PCS)
- Documentation covering the scenario, how the model works, how to read the results and how to build your own model
- A walkthrough notebook and exercises
- Tests, so you can check your setup works
Get started
You need Python 3.10–3.13. Full instructions are in the repository’s README.
Disclaimer
This is a simplified and adapted version of Trase’s supply chain mapping code, designed for ease of use by the public. Orvessa, its companies and all of its data are fictional. Please note that the internal Trase code has differences. Also note that we do not use the same supply chain mapping methods between commodities and countries. Get an overview of Trase methodologies here: trase.earth/methodology
