www.dataforcanada.org

Check the current status of all our services here: 👉 status.dataforcanada.org

Data for Canada exists to bridge the gap between open data availability, resiliency, and usability. We curate, clean, and re-engineer high-value Canadian datasets into high-performance, analysis-ready formats for data engineers, researchers/scientists, developers, and systems.

Canada creates incredible amounts of open data, from foundational road networks to federal census statistics and orthoimagery. However, these datasets are often locked in legacy formats, fragmented portals, or structures that require significant engineering effort to normalize. For our target audience, the "time-to-insight" is often bottlenecked by data preparation.

Data Stability: Beyond technical barriers, open data can be ephemeral. Links break, portals are reorganized, and priorities shift, causing valuable datasets to vanish from the public web. This instability makes it risky to build long-term research or software on top of data providers.

We prioritize our work in a utilitarian manner, aiming to provide the greatest amount of good to the greatest amount of individuals. Our approach is guided by the principles of digital preservation and the need to keep public information accessible over the long term.

Our approach is guided by the following:

• Link rot in LIS literature: a 20-year study of web citation decay, recovery and preservation challenges
• Guidance on Assessing Readiness to Manage Data According to the Findable, Accessible, Interoperable and Reusable (FAIR) Principles
• Sustainability of Digital Formats: Planning for Library of Congress Collections
• Science Needs a Social Network for Sharing Big Data
• Cloud-Optimized Geospatial Formats Guide
• GC White Paper: Data Sovereignty and Public Cloud

We act as the transformation layer. We aggregate datasets with permissive licenses and process them into "digestible" standards optimized for modern downstream applications.

• For Data Engineers, Researchers/Scientists, and Developers: Skip the cleaning phase. Access normalized, documented data ready for analysis.
• For Systems: Standardized data structures designed to feed directly into pipelines, data warehouses, and downstream services.

Our Stewardship: Data for Canada takes ownership of the datasets we create, from start to finish. We ensure that data remains consistent and available, acting as a stable foundation for your work, and allowing for reliable analysis across time and space. By decoupling access from the original source, we ensure your pipelines don't break even if the upstream location changes or expires.

We adopt an open-source first approach, while supporting proprietary solutions to the best of our ability to ensure maximum accessibility. We target the latest versions of these software packages (e.g., modern GDAL/OGR) to leverage the newest improvements.

Our data is optimized for:

See our processing pipeline in action. View samples and documentation for our current priority processes:

• Statistical Products: Census data and other quantitative datasets.
• Foundation: Core geospatial layers including address point, road networks, and buildings.
• Orthoimagery: High-resolution orthoimagery.

Note: The data sources in the diagram below are prioritized from left to right, reflecting our current focus on processing high-value statistical, foundational, and orthoimagery datasets first.

We are actively looking for new members and partners to help shape this project.

To support data sovereignty, safeguard against data loss, and improve local access speeds, we are currently seeking selective mirroring in Canada. See our Infrastructure.

We are looking for academic institutions, research organizations, or infrastructure partners interested in hosting mirrors of specific, high-value dataset subsets. If you have bandwidth and storage capacity to spare for the Canadian open data ecosystem, please contact us.

Right now, we primarily need feedback on file naming convention, our datasets and their underlying processes, and the infrastructure used to generate them. If you have thoughts on data quality, format optimization, or pipeline improvements, we want to hear from you.

• Discussions: Head over to #dataforcanada:matrix.org to chat, or go to the individual process GitHub repos to comment on specific issues.

This project is licensed under the MIT License.