Skip to content

README.md

Latest commit

 

History

History
52 lines (41 loc) · 2.54 KB

README.md

File metadata and controls

52 lines (41 loc) · 2.54 KB
Raw

Data Model & Build Pipeline

The library is data-first: every case study is a single JSON file in data/studies/, and all other data files are generated from those sources by the build pipeline. Never hand-edit the generated files.

Source of truth — data/studies/<slug>.json

Field Type Required Notes
id number ✓ Stable, unique. Controls default ordering.
slug string ✓ Unique, kebab-case. Must equal the filename (<slug>.json).
title string ✓ Display title.
country string ✓ Primary country (or Global).
region string ✓ One of the seven World-Bank-style regions.
topic string ✓ One of the ten topic labels.
year string ✓ e.g. 2007–present, 1990s–2010.
summary string ✓ 1–2 sentence card blurb (≥ 40 chars).
keyData array – { "label", "value" } pairs shown in the study sidebar.
content string ✓ Markdown body. Section headings drive the reader UI.

Recognised content section headings

The reader parses these headings to build the evidence card, references list, and study framework:

  • ## Results & Evidence / ## Impact / ## Outcomes — mined for quantified findings.
  • ## Key Lessons / ## Lessons for Policymakers — the "How to think about this" panel.
  • ## References / ## Sources / ## Bibliography — the citation list (drives the evidence tier).
  • ## Further Reading — supplementary reading list.
  • ## Discussion Questions — teaching prompts.

Generated artefacts (do not edit)

File Produced by Used by
data/master-list.json npm run build Card grid, filters, related studies. Enriched with readingMinutes, words, refs, evidence.
data/search-index.json npm run build (Reserved) lightweight client search index.
data/stats.json npm run build The Insights dashboard on the home page.
sitemap.xml npm run build Search-engine discovery (one URL per study).

Commands

npm run build      # regenerate all derived artefacts
npm run validate   # CI check: validate schema + fail if artefacts drift
npm run dev        # zero-dependency static server on :8000

Adding or editing a study is a two-step flow: edit the JSON in data/studies/, then run npm run build and commit the regenerated files. CI (npm run validate) fails the PR if you forget.