Skip to content

Tutorial: Thermal Stability Tracking

  • Data: oil/chips synthetic (or your own heating series)
  • Protocol: examples/protocols/EdibleOil_Heating_Stability_v1.yml
  • Outputs: Heating trend tables/plots, stability rankings, interpretation tables.

Tutorial – Thermal stability tracking

What this tutorial covers

  • Problem: Thermal degradation across heating stages; find heating-sensitive markers.
  • Dataset: examples/data/oils.csv (or chips) with heating_stage.
  • Protocol: examples/protocols/oil_heating.yaml (trend analysis, monotonicity).

Why it matters (theory)

Heating alters oxidation/unsaturation signatures. Trend analysis (linear slopes, Spearman ρ) shows whether ratios increase/decrease with heating. See rq_engine_theory.md.

CLI workflow

foodspec-run-protocol \
  --input examples/data/oils.csv \
  --protocol examples/protocols/oil_heating.yaml \
  --output-dir runs/oil_heating_demo

foodspec-publish runs/oil_heating_demo/<timestamp> --fig-limit 6
Open figures/trend_* for ratio vs heating stage; tables/heating_trend_summary.csv for slopes/ρ/p-values/monotonicity.

Example figures (from run bundle)

Heating trends

How to read the results

  • Slopes (p < 0.05 FDR): heating-sensitive markers.
  • Spearman ρ: near ±1 indicates monotonic change.
  • Stability ranking: lower abs(slope) = more stable spectral fingerprint.
  • Choose markers with consistent trend direction across replicates/batches for QA.