The breed dataset
Everything the site scores, in one file. BreedQuest holds around 5,400 rated values across 458 dog and cat breeds — temperament, care demands, energy, health — and until now they existed only spread across 458 separate pages. The table below is the same data, flat, generated on every build so it cannot drift from what the pages say.
What is in it
One row per breed, 36 columns. The identity columns are slug, name, species and group, plus the FCI number and group where the breed is FCI-registered and whether the AKC recognises it. Then four descriptive facts — size, weight, life expectancy and coat — as they appear on the breed page.
The scores are the point. Eleven trait columns carry the 1–5 ratings the breed pages display, and six more carry the 1–10 ratings the app uses. They measure different things on different scales and are deliberately not merged.
The signature trait, and why there are two of them
Every breed carries exactly twelve rated traits, but only eleven of a dog's are fixed. The twelfth is chosen per breed from 48 possibilities — Herding instinct, Snoring, Lapdog delusion, Water rescue instinct, Escape artistry — and it is frequently the most interesting number on the page. Cats have their own fixed twelve and share only eight labels with dogs.
Rather than flatten that away to make the grid rectangular, the export gives it two columns of its own: signature_trait and signature_score. Two breeds need a second pair, so there is one — the Pembroke Welsh Corgi carries Barking / alert plus Herding instinct, and the Siberian Husky carries Vocal / howling plus Escape artistry. A rectangular twelve-column grid would have silently dropped 364 breed-specific scores.
Where the numbers come from
The descriptive facts — size, weight, lifespan, coat, group — are read from breed standards, with the FCI and AKC as the primary references and the national club where a breed is theirs. Those are checkable, and the breed page cites its sources.
The trait scores are editorial ratings, not measurements. They are our reading of the breed standard and the working literature, applied consistently across the library so that comparisons between breeds mean something, and they are reviewed rather than generated. Treat them as a well-founded opinion with a consistent scale — which is what they are — and not as survey data. Where a value is derived rather than sourced, the method is documented: the playfulness score, for instance, is a published formula over traits the breed already carries, with editorial corrections that each cite a standard.
Using it
Both files are plain HTTP downloads with no key and no rate limit. The CSV is UTF-8 with a header row. The JSON is the registry the site itself reads.
We have not put a reuse licence on this data, so no rights are granted beyond reading it and citing it. If you want to build something on it, that is a conversation we are glad to have — write to [email protected] and tell us what you have in mind. Citing it needs no permission: BreedQuest, breed dataset, with the date you downloaded it.
The columns
slug, name, species | the breed's identity and whether it is a dog or a cat |
group | our group key — hound, terrier, shorthair… |
fci_number, fci_group | FCI nomenclature, blank for cats and for unregistered breeds |
akc | yes / no |
size, weight, life_expectancy, coat | the descriptive facts, as printed on the breed page |
web_* | eleven trait scores, 1–5, the ones the breed pages display |
signature_trait, signature_score | the breed-specific twelfth trait, and its score |
signature_trait_2, signature_score_2 | a second one, for the two breeds that carry two |
app_* | six trait scores, 1–10, the scale the app uses |
Cite as: BreedQuest. Breed dataset. breedquest.com/data/