Almagest

A data intelligence platform for Canadian government open data.

Ptolemy's Almagest compiled the known cosmos into a single coherent treatise. This project compiles Canadian government data into a single intelligible model.

The problem

Canada's federal government publishes roughly 30,000 datasets through open.canada.ca, covering on the order of $450 billion in annual federal spending: grants, contracts, lobbying disclosures, departmental performance, and the long tail of agency-specific data.

The data is freely available and almost entirely unusable for analysis. Formats shift across years. Entities are not resolved across sources: a single grant recipient appears as "University of Toronto," "U of T," and "University of Toronto, Faculty of Medicine" in the same dataset, same year. Schemas drift silently as columns are renamed or added.

The result is that a question like how much total federal funding has a given Canadian university received across all departments in the last decade takes weeks of manual data wrangling, if it gets answered at all.

The approach

Almagest applies the same two-layer pattern Palantir Foundry uses for enterprise data, applied to public data. The raw layer holds every CKAN dataset on open.canada.ca, content-addressed and version-tracked. The ontology layer holds a small set of typed canonical objects — Recipient, Department, Program, Grant, Contract — with resolved entities and full lineage back to source.

Apps and AI operate exclusively over the ontology. The raw layer is reproducible by any competent team in a few months. The ontology — specifically the entity resolution that makes "RBC," "Royal Bank of Canada," and "Royal Bank of Canada (Toronto)" cluster correctly across millions of rows — is the work that compounds.

Raw layer

Every CKAN dataset on open.canada.ca, content-addressed and version-tracked. Comprehensive coverage, unmodeled.

Ontology layer

Typed canonical objects with resolved entities and full lineage. Where apps and AI actually live.

Live tool

What exists today

Before Almagest, the same problem space produced OpenBSIS — a working tool for Canadian bank regulatory disclosures, live at openbsis.com. It's how we got into this problem space in the first place: a single corner of Canadian public data, made legible.

Almagest is the broader thesis — that the same approach generalizes across federal open data. Active development on OpenBSIS is parked while the platform foundation is built.

About

Almagest is pre-launch. The intended audience is journalists, researchers, and civic-tech — people for whom Canadian public-data infrastructure is a load-bearing tool rather than an internal asset.