DefenseFood scores every commodity, destination, origin lane in your data by combining RASFF hazard alerts, UN Comtrade bilateral trade, and FAOSTAT food balance sheets. You get a ranked priority queue. Click any row to see how the score was built, term by term.
Designed for EU food safety planners, sampling teams, and food fraud researchers. Every headline number maps to a published equation in the framework, with the inputs and value bands shown alongside the score.
7
Framework sections
Sections 2 through 7 of the v1.0 framework
22
Documented metrics
Each with a formula, scale bands, and inputs
6
Hazard families
Biological, mycotoxins, pesticides, metals, other chemical, regulatory
4
RASFF roles
notifier, distribution, follow-up, attention
What you can do here
Sort lanes by combined vulnerability (CVS) and filter by destination, origin, hazard family, or market presence. The default view hides informational only RASFF mentions, so the queue reflects lanes that are actually on EU markets per the SOPs.
Browse corridorsCountry pages show inbound exposure (ACEP) and outbound propagation (ORPS), split by RASFF role. The exposure graph colours confirmed market lanes by hazard intensity and distinguishes them from informational mentions.
Open the network viewEvery metric has its own glossary card with the blueprint formula, a plain English rewrite, scale bands, and which inputs went in. Open any lane to see how its CVS was built, with each amplifier term struck through when the input is missing.
See the glossaryHow it works
RASFF Window exports for hazard alerts. UN Comtrade for bilateral trade. FAOSTAT for production and consumption. The merged trade corpus is prepared ahead of time so requests stay fast.
A Rust engine evaluates the seven sections of the framework on every corridor at startup: dependency, consumption demand, hazard intensity, trade flow anomalies, network propagation, and composite scoring.
Each score on the dashboard links back to its bands, its formula, and the inputs that drove it. When data is partial, the system says so instead of hiding the gap.
What is in the engine
Sections 2 through 7 of the v1.0 mathematical framework are computed at startup. The methodology view documents the formula, inputs, scale bands, and Rust function for each metric.
IDR, OCS, BDI, HHI, SCI: how reliant a destination is on an origin for a given commodity.
PCC, CRS, DIS: how culturally entrenched the commodity is in the destination's diet.
HIS, HDI, DGI: severity weighted, time decayed alert pressure with detection gap diagnostics.
Unit value z scores, volume anomalies, mirror trade discrepancies, year on year concentration shifts.
ORPS, ACEP, empirical hazard probability. Role aware aggregation following Pan et al. (2025).
Hybrid CVS with masked amplifier terms (Slice E1), neutral CRS fallback, percentile re anchored bands.