The Evidence Atlas · a general engine, shown here on one problem
Atlas engine v3.14 · Catalogue data v3.12i · sources checked against primary reports · Aug 2026
Take a problem, follow it through the food system, and see what the evidence actually shows at each stage — and how far it reaches.
This is not an atlas about aflatoxin. It is a way of laying any food-system problem along its value chain, mapping the practices tried at each stage against what was measured and what it achieved — positive, mixed, not achieved, or not yet determined — with every figure keyed to a primary source. Aflatoxin is Demonstration 1: one fully-worked problem drawn from the catalogue, chosen because it stress-tests the whole chain from field to plate.
Problem→ value-chain stage→ practice→ what was measured→ result→ the outcome that matters
The catalogue holds 96 records across all seven problem classes (Data Descriptor v3.12i). The two live demonstrations wire a subset of two of those classes; the other five are populated in the catalogue but not yet built into the atlas view.
How to read this

Each demonstration follows one problem along the food system: stage → practice → what was measured → did it work? → was the endpoint reached? → how far the evidence reached → primary source.

  • “Did it work?” and “endpoint reached?” are different questions — a practice can do the first without the second.
  • Evidence reach is the furthest outcome the evidence actually established, not a score: the positions are named endpoints, not points on a 0–100 scale.
  • Depth is not quality. Reaching a further endpoint means the evidence travelled further, not that the study was stronger than one with a rigorously measured nearer endpoint.
  • “Not determined” means no conclusion could be drawn from the available evidence — not that nothing happened.
  • Position on the chain is not a ranking. Upstream controls usually have more leverage on the problem, but a downstream check can be the decisive step that keeps unsafe food off the plate.

Countries with catalogue evidence7 countries · click a marker
Where each practice was tested. The evidence itself — what was measured and whether the endpoint was reached — is in the panel on the right.
✓ Positive ~ Mixed / conditional ✗ Not achieved ? Not determined
“Not determined” = no conclusion from the available evidence, not “nothing happened”.