An interactive drill for the FAIR risk taxonomy: study the decomposition tree, then rebuild it from memory — place all 13 components in the right positions, assign the right unit to each one, and match all 22 testable definitions.
Live: https://xnasusx.github.io/fair-model-study/
-
Study mode — the complete FAIR decomposition tree, from Risk down to Secondary Loss Magnitude, with the unit each factor carries: dollars (
$), probability (%), or frequency (#). -
Practice mode — the tree comes back empty. Pick a component from the shuffled pool, tap where it belongs, then assign its unit before moving on. Placement accuracy and unit accuracy are scored separately, so you can see which half you actually know.
-
Definitions quiz — 22 items in one pool: the 13 model components, the 6 forms of loss, and the 3 Probability of Action sub-factors. Match each name to its definition; colour-coded by category.
-
Worked example — load a real, source-backed scenario onto the study tree and watch how little of it fills in. A published loss study reports an annualised event frequency and a per-event loss: in FAIR terms, Loss Event Frequency and Loss Magnitude, and nothing else. Two of thirteen nodes. Nobody measures contact frequency or resistance strength across a population, so the eleven empty nodes are not missing data — they are the analysis, which is the reason the decomposition is worth knowing cold. Each filled node shows the sources behind it, with their stated limitations.
The tree scales itself to fit the viewport, so the full 13-node model stays readable on a phone.
Single self-contained index.html — React 18 via UMD CDN, no build step, no dependencies.
Styled to match the palette and type system of my portfolio: powder
rose surfaces, warm ink, rose accent, Fraunces for display and Inter for UI.
python -m http.server 8000Then open http://localhost:8000. The worked example fetches the shard corpus over HTTPS at load; if that is unreachable the panel says so and both drills work exactly as before, since nothing in the study or practice modes depends on it.
Built as a training exercise — a study aid for learning the FAIR taxonomy, and for learning how this kind of interaction is put together.
The taxonomy follows the FAIR Model Standard Artifact Version 3.0 (January 2025) published by the FAIR Institute. This is a study aid, not a substitute for official FAIR training material. The FAIR Model™ is a trademark of the FAIR Institute.
Worked-example data comes from risk-benchmarks, which derives it from RiskShard by raviaxo, AGPL-3.0. The mapping of a shard onto FAIR nodes is this project's own, and it is deliberately shallow: the shard's frequency triangle is Loss Event Frequency and its impact triangle is Loss Magnitude, with no attempt to derive the factors beneath them, because nothing in the source supports that derivation.
Copyright (c) 2026 Susan Shepard.
GNU AGPL v3 or later. If you modify this and run it as a network service, the AGPL requires you to offer your users the modified source under the same terms.