The Siren Seafood Guide

Methodology: how we build species profiles

Every species profile on this site is a join of two public datasets. No portion of the displayed text is machine-generated or rewritten — if a paragraph appears on this site under NOAA’s name, NOAA wrote it.

Primary source — NOAA Fisheries Sustainable Seafood Directory

We scrape the species pages listed at NOAA’s sustainable-seafood directory. Each species page contributes the following fields verbatim: taste, texture, health benefits, availability, source, population status, appearance, biology, where they live, fishery management, harvest, population, fishing rate, habitat impacts, bycatch, and the species illustration. Sustainability status sentences like “U.S. wild-caught … is a smart seafood choice” are quoted directly from NOAA.

Supplementary source — USDA FoodData Central (SR Legacy)

Nutrition data comes from USDA FoodData Central, SR Legacy release. We import the bulk CSV, filter to the Finfish & Shellfish Products category (food category 15), and extract 24 nutrient columns — protein, fats, cholesterol, sodium, EPA/DHA/ALA omega-3s, vitamin D, B12, selenium, and major minerals. SR Legacy is a frozen release and will not update; this site therefore does not need a live sync.

How species are joined

NOAA uses common names (e.g. “Alaska Pollock”); USDA uses descriptive text (e.g. “Fish, pollock, Alaska, raw”). We tokenize both, strip geographic adjectives (“Atlantic”, “Pacific”, “Alaska”, “North”) and preparation words (“raw”, “cooked”, “mixed species”), and keep only the species-identifying tokens. A USDA row is accepted when at least one species token overlaps; among candidates, the highest token-set-ratio wins. 111 of 121 species match cleanly. When USDA has no entry for a species, the page falls back to NOAA’s own nutrition-facts text block.

What we do not do

  • We do not rewrite NOAA’s text with AI. The paragraphs are displayed verbatim.
  • We do not show placeholder or rounded-to-zero values. If a nutrient field is missing for a species, that row is hidden on the page.
  • We do not hot-link images. Every NOAA illustration is mirrored locally at the time of import.

Imagery credits

Ambient and editorial photography on this site is public domain, sourced from Wikimedia Commons and the NOAA Photo Library on Flickr. Individual files: Salmon troller, Lifting salmon from a hold, BCF Research Vessel ALBATROSS IV at Woods Hole pier, Optical effect march sunset — NOAA, and Oysters served on ice, with lemon and parsley. No attribution is legally required for these files, but we name them here as part of our editorial source policy. Per-species scientific illustrations are by NOAA Fisheries/Jack Hornady and colleagues, reproduced locally from the NOAA species directory.