SynthPopCan prepared model: Toronto CMA 2021 broad linked package

Publishable-candidate linked household/person model trained from the 2021 Census hierarchical PUMF for Toronto CMA (CMA 535). Geography: Toronto CMA (CMA 535)Census vintage: 2021 CensusConditioning columns: PR, household_size, TENURPackage version: v1.0.0-rights.1 Source attribution. Adapted from Statistics Canada, Hierarchical File, Census of Population: Public Use Microdata Files, Census Year 2021 (98M0001X2021002), 2021. This does not constitute an endorsement by Statistics Canada of this product. Layer-specific rights. Zenodo's compatibility label is Other (Open). The machine-readable contract is synthpopcan-prepared-model-licensing-v1. The licence layers are cumulative, not alternatives. CC BY 4.0 does not license, replace, or supersede Statistics Canada Information or rights governed by the Statistics Canada Open Licence. CC BY 4.0 is offered only for rights Darcy Quesnel owns or controls in the named original material, and only to the extent that material is protected by applicable law. The Creative Commons Attribution 4.0 International is the authored-layer licence. Named original materials: original selection, original organization, original documentation, original schema representation, original model representation. Excluded material: Statistics Canada Information, source classifications and labels, facts and factual data, numeric results not protected by applicable law. The source file is released under the Statistics Canada Open Licence; retain its required attribution above and comply with its accuracy, non-identification, non-misrepresentation, and no-endorsement conditions. Project rights-policy status: accepted. External legal review: not-obtained. Disclosure review. No raw rows or source identifiers. Review status: publishable candidate; human review still required. Passing SynthPopCan's disclosure-risk checks is a project-level screen, not legal anonymization or Statistics Canada endorsement. Known limitations. Broad 2021 PUMF model; not calibrated to small-area controls. Generated source codes require field metadata for interpretation. Some sparse geographies use privacy-safe CART models instead of direct conditional frequencies. Generation guidance. Use durable Python-backed web runs or the CLI; review memory, disk, and output scale before large generation. Corrected package integrity. Compressed 1,636,646 bytes, SHA-256 37a46b9fd59a60b068c206d9ebc288bfc00b747394c8ce4d570647cc8a71d40c. Uncompressed 100,650,715 bytes, SHA-256 45bddb0ae7520ea256c5622d32d73799c4be2711b776f22869d0714ba70539c6. Generated populations are synthetic artifacts. They are not real Census records and must not be presented as confidential Statistics Canada information. Corrected package. These candidate bytes embed the validated synthpopcan-prepared-model-licensing-v1 contract at top-level licensing.