Live demo: mhofman11.github.io/dat
A 5-minute diagnostic that shows where daily life is making dietary adherence harder.
Most diet failures aren't knowledge problems. People know what to eat. The problem is that the environment they live in, the biological signals their body sends, and the cognitive bandwidth they have left at decision time create friction that makes consistent follow-through harder than it should be.
DAT maps that friction across three axes and tells you which one is dominating your adherence outcomes — so interventions target the actual constraint instead of adding more pressure to a system that's already overloaded.
DAT is built on a five-dimension framework called PEBCS — Plan, Environmental, Biological, Cognitive, and Social — drawn from adherence research including Dansinger et al. (JAMA, 2005), Johnston et al. (JAMA, 2014), and Ge et al. (BMJ, 2020). The consistent finding across this literature: adherence predicts outcomes better than diet type.
Of those five dimensions, DAT directly measures three:
-
Environmental Friction (E) — What does your physical world actually allow? Food access, kitchen setup, budget, proximity to options, work schedule. The gap between what you'd eat and what's realistically available.
-
Biological Friction (B) — What is your body doing to you? Sleep debt, energy crashes, hunger volatility, circadian disruption. The physiological signals that override intention before conscious choice occurs.
-
Cognitive Friction (C) — How much bandwidth do you actually have? Stress load, schedule unpredictability, decision fatigue. By the time food decisions come up, there's often nothing left in the tank.
Plan quality (P) is not directly scored but is flagged conditionally — when biological or cognitive scores are high, the prompt layer screens for whether the diet structure itself may be contributing. Social friction (S) is acknowledged but excluded from scoring; it requires diary-based assessment methods that don't fit a 5-minute intake, and the tool assumes conditions of ordinary adult autonomy over food decisions.
The assessment collects structured inputs across the three measured axes and produces:
-
A friction profile — three independent scores (0–100) visualized on a triangle, with a composite stability score calculated via harmonic mean. The harmonic mean naturally penalizes single-axis failure, reflecting how one unstable dimension can destabilize the whole system.
-
An archetype classification — the dominant axis determines which of 18 behavioral archetypes best describes the user's friction pattern, routing interpretation toward the actual constraint.
-
Two AI-ready outputs:
- An AI starter prompt — paste into ChatGPT or Claude to start a coaching conversation that already knows your friction profile, dominant axis, and intervention priorities.
- A Claude Skill file (.md download) — a reusable coaching skill that can be loaded into Claude for ongoing, context-aware dietary guidance.
Both outputs include plan-quality checkpoint logic (triggered when B or C is dominant) and a three-path intervention framework: reducible friction (fix at source), structural friction (design around it), and plan-generated friction (flag the diet itself).
Single-page application. HTML, CSS, and vanilla JavaScript. No frameworks, no dependencies, no server. All scoring logic is deterministic, transparent, and runs client-side.
Architecture:
- Deterministic scoring engine with independent axis calculations
- Harmonic mean composite scoring
- Rule-based archetype classification (18 archetypes across 3 axes)
- Structured prompt and skill template generation from assessment output
- Responsive UI with progressive disclosure across 6 steps
Functional prototype. Conceptual architecture validated through iterative review. Scoring engine stable. Ready for live user testing to calibrate question weights, archetype thresholds, and intervention effectiveness.
Developed from applied fitness coaching and behavior change experience. The consistent observation: outcomes are better predicted by environmental and cognitive constraints than by nutritional sophistication. DAT formalizes that pattern into a reusable diagnostic.
Michael Hofman