As part of accomplishing our goal of bringing together the USEEIO and CEDA into a single global, open MRIO (Multi-Regional Input-Output) model, we drafted a technical architecture vision for the development of this new model.

Today, we are publishing this vision, available to download here.

This document explores how we should structure our technical tools, systems, and workflows. The key points to highlight are:

1. Cornerstone will develop in Python
2. Core Cornerstone code should be stored in a monorepo along with a supporting cloud storage system for datasets, allowing easy sharing of utilities, tools, standards, and data
3. We should think of our environmentally-extended input-output model (EEIO) (and future non-EEIO) data development as building an Extract-Load-Transform (ETL) pipeline, with clear code and clear workflow boundaries: extracting, transforming (with the latest methodologies), and loading/publishing data.

We welcome feedback and reactions to our vision via github.