About Deep Data
A search engine for Canada's open government data.
Our mission
Canada's federal and provincial governments publish millions of records every year — grants, contracts, patents, charity filings, regulator registries, import data — under open licences. Each source lives in its own portal, format, and update cadence. Useful in isolation; powerful only when combined.
Deep Data is an independent private project that aggregates, normalizes, and indexes these public datasets into a single searchable corpus. We unify bilingual names, deduplicate entities across sources, preserve originals alongside normalized forms, and link every row back to the government source it came from. The goal: make Canada's open business data actually usable for research, due diligence, journalism, and market intelligence.
We are not affiliated with, endorsed by, or authorized by the Government of Canada or any federal or provincial agency. All source data remains attributable to its original publisher under their respective licences.
Data sources
Every record in Deep Data traces back to a public government dataset:
- Statistics Canada — Open Database of Businesses
- Open Government — Federal grants, contracts, and departmental filings
- Canadian Intellectual Property Office (CIPO) — Patents and trademarks
- Office of the Superintendent of Financial Institutions — OSFI regulated entities and pension plans
- Financial Transactions and Reports Analysis Centre of Canada — FINTRAC MSB registry
- Canada Revenue Agency — Registered charities (T3010 filings)
- Environment and Climate Change Canada — National Pollutant Release Inventory (NPRI)
- Registraire des entreprises du Québec via Données Québec (CC BY 4.0)
- Open data portals for Toronto, Vancouver, Calgary, and Edmonton
Each profile page links back to the specific sources that contributed to it.
How the data is updated
Source datasets are ingested on a regular cadence — most federal datasets refresh weekly, CIPO patent data monthly, provincial registries on their own schedules. Each import runs in a single transaction, is logged to an audit trail with per-row lineage, and can be rolled back surgically if a source publishes corrupted data.
When a company's core details change — legal name, headquarters address, business number — we archive the previous version with a timestamp range, so you can query the state of any company as of any historical date. Nothing is silently overwritten.
During normalization we never discard original values. Company suffixes (INC, LTD, LTÉE), bilingual name pairs, operating-as designations, and jurisdiction brackets are all extracted into structured fields; the raw source form is preserved. Matching a record to a real-world entity is a confidence-scored operation, not a hard merge.
Contact
Questions, data corrections, licensing inquiries, or press — [email protected].
If you represent a company profiled on Deep Data and want to update or remove information, include your business number or website in the message and we'll respond within a few business days.