xiaoyang111222/cn-math-taxonomy

The problem

Anyone building study tools for Chinese K-12 students hits the same wall: there is no machine-readable map of what depends on what. Textbooks and curriculum standards are PDFs full of prose; the prerequisite chain behind "completing the square" only exists in teachers' heads. This project turns that into data: 1,644 knowledge points, 2,403 prerequisite edges and 6,723 questions spanning primary-to-high-school math, English and physics, plus a diagnostic tool that walks errors back along those edges to an earlier concept.

What's different

Most curriculum datasets are flat lists of topics or exam-question banks tagged by chapter. Here the value is the edge set — hard vs. soft prerequisites, cross-stage links from junior-high back to primary arithmetic — which enables the backtracking diagnostic rather than simple tagging. It is also unusually explicit about provenance: which textbook table of contents came from the official smart-education platform, which from third-party sites, and which are marked 待核. Compared with commercial adaptive-learning products, it is open data and inspectable, not a black box.

Who it's for

Edtech developers who need a prerequisite graph or question bank under a Chinese curriculum; researchers studying knowledge tracing; teachers who want to see the dependency chain for a topic. It is not for anyone needing classroom-ready, teacher-vetted content: questions for high-school math and all English have not been blind-solved, and item difficulty or discrimination data is absent.

Cost of adoption

Python plus a static JSON data format (topics + dependencies, borrowed from os-taxonomy). The README is long but mostly a provenance ledger — worth reading before trusting any edge. The license badge says CC BY-NC 4.0, so commercial use is restricted; the repository's LICENSE metadata does not confirm this, so verify before shipping. The top risk is quality: the author states plainly that this is an AI-generated first draft with no teacher review, that some scope (integrated math, some textbook volumes) follows unverified third-party tables of contents, and that English lacks listening/speaking items. Budget for human review, not just integration.

Verdict

This is a six-day-old project with roughly 3,200 stars and about 500 more per day, so the attention is a spike, not a track record, and it has not appeared on the daily trending board. The underlying artifact — a curriculum-aligned prerequisite graph — is genuinely scarce and hard to assemble, which explains the interest. But the author's own disclaimer that nothing has been teacher-reviewed should be taken at face value: treat this as a strong draft to fork and audit, not a drop-in content source. It is worth an afternoon if you need the graph structure; skip it if you need verified items.