05 / Learn / Canadian citizenship study
A study system that knows its sources.
Official Canadian citizenship material, turned into a focused Anki study deck—with source checks, a prioritized core, and visible reminders when facts can change.
127 more retained for later
Usable deck · designed for ongoing verification
- cards in the deck
- 441
- cards in the study core
- 314
- cards flagged for review
- 20
01 / The brief
Build something dependable enough for a real exam.
The project converts the official Discover Canada guide and current IRCC test pages into focused recall prompts. AI helped with research, drafting, and implementation, but it was never allowed to serve as the source of a fact: every curricular claim had to trace back to fetched government material.
02 / The workflow
Separate generation from verification.
A reproducible fetcher collects the source pages, the card data stays in a reviewable Python structure, and a build script packages the result for Anki with stable identifiers and chapter tags. A separate audit compares numeric literals and proper-name candidates with the fetched text; it is a useful consistency check, not a substitute for editorial review.
03 / The judgment
Optimize the study plan, not the card count.
The full deck keeps 441 cards, but 127 lower-yield cards ship suspended so a new learner begins with a 314-card core. The selection draws on the governing regulation, IRCC's own study questions, and a public practice bank. Nothing is deleted, so the learner can restore any topic instead of accepting a hidden editorial decision.
04 / Source drift
Preserve the exam answer while making stale information visible.
The official guide contains figures and terminology that no longer match present-day Canada. Twenty cards carry a visible review note: the answer preserves what the guide teaches for the exam, while the note explains that the real-world fact has changed or must be checked again. That distinction turns uncertainty into part of the interface rather than a footnote in the process.