Design decisions¶
| # | Topic | Decision | Reason |
|---|---|---|---|
| D1 | Licences | CC BY 4.0 for content, MIT for code; data keeps its own licence | Reuse with credit; CDC and NGOs can adopt without share-alike friction |
| D2 | Languages | English first, Nepali by a teacher-editor, released together per unit | Technical content is faster to write in English; Nepali is needed for CDC adoption |
| D3 | Dashboards | Streamlit; Django only as an optional Grade 12 stretch | Light on school PCs; stays focused on data |
| D4 | Programming order | SQL in Grade 9, a short first-Python taste in Grade 9, main Python in Grade 10 | Matches the optional CS syllabus as reported |
| D5 | Database | SQLite in Grades 9–11; MySQL/PostgreSQL as concepts in 11–12 | No install, runs in the browser, same SQL |
| D6 | Health data | No DHS data is shipped. A clearly labelled synthetic table is used instead | DHS terms forbid redistribution "directly or within any tool/dashboard" |
| D7 | Real vs synthetic data | Real data is fetched from the original publisher by script; only small synthetic tables are shipped, labelled SYNTHETIC |
Licence safety; no invented numbers presented as real |
| D8 | Data licences | Share-alike data (CC BY-SA, ODbL) stays in datasets/ and is not mixed into CC BY text |
Keeps CC BY content clean |
| D9 | Tooling | MkDocs Material + JupyterLite + GitHub Pages + GitHub Actions; Colab as secondary | Contributor-friendly, offline-capable, free |
| D10 | Offline first | JupyterLite primary; school USB bundle (tools/build_offline_bundle.py) |
Weak connectivity in many schools |
| D11 | Name and branding | "Data Science Nepal / डेटा विज्ञान"; repo data-science-nepal |
Neutral and searchable |
| D12 | Hosting | GitHub account rcgiri-physics; move to an organisation once there are two or more maintainers |
Simple start |
| D13 | Release scope | Grade 8 Unit 1 and Grade 9 Unit 3 are the v0.1 release candidate; other units are available for review and are not recommended for classroom pilots until reviewed | Review the vertical slice before scaling |
| D14 | Pilots | 3–5 schools: urban and rural, public and private, with and without internet; first units G8 U1 and G9 U3 | Tests both tracks |
| D15 | Co-maintainers | One teacher, one statistician or university contact, one Nepali-language editor | Review quality and CDC credibility |
| D16 | Releases | v0.x until pilot evidence exists; v1.0 afterwards |
Avoids claiming unproven impact |
| D17 | Book source | The book is assembled from the repository's markdown: chapters stay in curriculum/, book-only parts in book/ | One copy of each text; site, notebooks and book never drift apart |
| D18 | Editions | Student edition (chapters, ★ answers) and teacher edition (adds teacher guides, all answers, teaching handbook) from the same sources | Self-study and teaching needs differ |
| D19 | PDF engine | python-markdown → HTML + print CSS → headless Chrome/Edge; page numbers in the contents via a second pass with pypdf | No LaTeX setup; Devanagari via system fonts; runs on GitHub's Ubuntu runners |
| D20 | Real data in exercises | Exercises link to publishers (RD-01…RD-20) and always have an offline fallback on synthetic tables; D7/D8 unchanged | Real data without re-hosting or licence risk |
Dataset licence evidence: alignment/dataset-licence-register.md.