Toronto’s Questions — the data

The package’s own machine-readable corpora, published as plain CSV so anyone can check the work, re-cut it, or build on it. These are the same files the pages are built from — no reduced or cleaned-up public edition. Free to use, with or without attribution.

Everything else this project publishes as open data is on the open data page; the code and the source markdown behind these pages are in the commons repository.

question-harvest.csv

The question corpus — every research question in the package, machine-harvested from the project’s own question banks and the register’s gap cells, each row carrying the file and ID it came from.

Download question-harvest.csv — 1,234 rows, 5 columns: source_program, source_file, source_id, source_kind, question

shelf-map.csv

The artifact map — which published research artifact serves which problem, many-to-many (one topic slug can serve several problems, and one problem draws on several slugs).

Download shelf-map.csv — 192 rows, 3 columns: slug, problem_id, strength

shelf-stems.csv

The raw stem inventory the artifact map was built from — the unreduced list, published so the mapping above can be checked rather than taken on trust.

Download shelf-stems.csv — 296 rows, 2 columns: stem, kinds

Found an error in the data? Tell us — corrections are published, not quietly patched.