Methodology
How OconEco produces each published number, with the audit trail to back it.
Two features anchor the methodology section: the Operational Guidelines (how we handle specific methodological domains) and NUSAP (how much to trust each number). Neither is front-page material, but they signal to institutional reviewers that this is peer-grade work.
The eight sub-pages
- Assessing National Statistics — NUSAP beside the IMF’s DQAF and Open Data Watch’s ODIN: who grades the producer, the release and the number, and where FAND sits.
- How Precise Is the Q-Matrix? — a place’s quintile is a band, not a point: what bounds the error, what does not, and a test against local figures.
- Intractable Index Number Problems — why no index keeps V = P × Q meaningful from wheat to food, across years and places; what each system gives up; and why FAND publishes value.
- Operational Guidelines (OG-001..OG-013) — 13 guidelines codifying methodological domains. Accordion: one-line summary visible, full text on expand.
- NUSAP — data quality — the Funtowicz & Ravetz framework, and how far FAND’s NUSAP-lite has taken each of its five tags.
- FAND’s switches — where FAND keeps two defensible choices side by side, and why.
- Source agency directory — every number links to its source agency, table, year, and URL. This is the directory of sources we cherry-pick from.
- Replication package — the 12-stage Python pipeline, the SHA-256 audit trail, and the public GitHub or gated download.
Why methodology gets its own section
Most data portals show you a number without telling you how much to trust it. FAND is built to show the quality metadata alongside every value. That is a differentiator institutional reviewers will notice — and a forcing function on us to not publish numbers we cannot defend.