Methodology

Scope

This page explains how PupWiki produces breed scores, cost estimates, fit scores, and tool outputs. We believe every number or recommendation should be explainable.

Breed trait data

Breed traits are derived from:

• Official breed standards from registry organizations (AKC, UKC).
• Veterinary references where breed-specific health traits are documented.
• Structured owner experience surveys (on the roadmap for future expansion).

Traits are expressed as ordinal levels (low, medium, high) rather than numeric scores to avoid false precision. Trait levels reflect general breed population tendencies, not guarantees about individual dogs.

Cost estimates

Cost estimates are based on:

• US national average pricing for food, routine veterinary care, flea/tick/heartworm prevention, and grooming.
• Breed-size cost adjustments (small, medium, large).
• Lifestyle tiers (budget, typical, premium) with documented multipliers.
• First-year estimates include one-time setup costs (supplies, initial vet visits, vaccines).

Cost estimates do not include: emergency veterinary care, major surgery, boarding, pet insurance, training classes, or city-specific cost adjustments.

Confidence levels

• High — Based on multiple Tier 1 sources, reviewed by domain expert.
• Medium — Based on Tier 1–2 sources, reasonable assumptions, no expert review yet.
• Low — Limited data, estimated from proxies, or broad assumptions.

Fit scores

The Breed Fit tool produces a 0–100 score based on a simple weighted model: household type, experience level, and daily alone-time tolerance. This is a starting-point estimate, not a scientific assessment. Individual dogs vary significantly from breed averages.

Limitations

• All breed data reflects population averages; individual dogs vary widely.
• Cost data uses US national averages; local costs may be significantly different.
• Health watchlist items are educational awareness, not veterinary guidance.
• No tool output should replace consultation with a licensed veterinarian or reputable breeder.

Update cadence

Cost models are reviewed annually. Breed trait data is reviewed when authoritative sources update. All pages show a "last reviewed" date.

Per-tool methodology

Each tool uses a defined, rule-based approach. The following sections describe how each tool produces its outputs, what assumptions are made, and what the outputs do and do not represent.

Compare Dog Breeds

Breed Compare displays trait values side by side for up to four breeds simultaneously. Visual advantage indicators (higher/lower/same) are derived by direct comparison of ordinal trait values from breed profile data. Health risk flags surface conditions documented in authoritative veterinary references for each breed. "Best for" chips reflect composite scoring against seven lifestyle scenarios:

• Apartment living — weighted by energy, size, barking tendency, apartment fit rating.
• Active family — weighted by energy, exercise needs, playfulness, family friendliness.
• First-time owner — weighted by training difficulty (inverted), affection, and independence (inverted).
• Allergy-friendly — weighted by shedding (inverted) and drooling (inverted).
• Watchdog — weighted by barking tendency, alertness, and protective instinct.
• Cold climate — weighted by cold tolerance and coat type.
• Travel-friendly — weighted by size, energy, adaptability, and alone-time tolerance.

Scenario scores are relative rankings across the breeds being compared, not absolute scores against all 70 breeds. A "winner" in a scenario simply means that breed scored highest within the current comparison set.

Dog Cost Calculator

The Cost Estimator calculates annual and first-year ownership cost ranges using:

• A base cost model with line items for food, routine vet care (2 visits/year), flea/tick/heartworm prevention, and grooming.
• Size-class multipliers (toy, small, medium, large, giant) applied to food and vet line items.
• Three lifestyle tiers (budget, typical, premium) with documented cost multipliers for each line item.
• First-year additions: supplies, spay/neuter, initial puppy vaccines, microchip.

Outputs are planning ranges only. They do not include pet insurance, boarding, emergency care, training, or city-level cost variation. Actual costs depend on location, individual dog health, and provider pricing. A breakdown chart shows which cost categories are largest. A scenario comparison table shows all three tiers side by side for planning purposes.

Dog Breed Selector Quiz

The Breed Fit tool produces a 0–100 score and a decision verdict (Great Fit, Good Fit, Fair Fit, Challenging Fit) for a specific breed against a set of user-provided lifestyle inputs. The weighted scoring model uses these factors:

• Household type (apartment/house with small yard/house with large yard/rural) — compared against the breed's apartment fit and size.
• Experience level (first-time/some experience/experienced) — compared against breed training difficulty.
• Daily alone time — compared against breed alone-time tolerance.
• Activity level — compared against breed energy and exercise needs.
• Presence of children — compared against breed family-friendliness and patience ratings.
• Budget factor — annual typical cost compared against user-specified budget tier, with a penalty applied if cost is substantially above budget.

Each factor contributes independently to the total score. The tool does not model interactions between factors. Sensitivity tips flag which factor(s) most affect the score. Alternative breed suggestions rank other breeds by score when the primary breed scores below 70.

Trait Finder

The Trait Finder is an inverse search: users specify desired trait levels across up to eight categories (energy, shedding, grooming, barking, trainability, good with children, good with other dogs, apartment fit) and the tool ranks all 70 breeds by how closely they match the specified preferences.

Match percentage is calculated as the proportion of selected trait preferences that a breed satisfies. An "exact match" on a trait means the breed's level equals the requested level. Trait data comes from breed profiles; match scoring is deterministic and rule-based. The "top 5 only" toggle restricts results to the highest-scoring breeds to reduce cognitive overload when many breeds match.

Training Plan Generator

The Training Plan Generator uses a rule-based, deterministic engine. It does not use AI or language models to produce outputs. Every training recommendation is derived from structured data files referencing AVSAB, AAHA, and Merck Veterinary Manual guidance.

The engine operates in the following sequence:

1. Safety gate evaluation — Seven gates check for bite history, aggression severity, fear/anxiety severity, separation panic, medical concerns, and aversive tool use. Any gate that triggers returns a safety-only result with professional or veterinary referral guidance instead of a training plan.
2. Life stage classification — Age in months maps to one of six life stages (young puppy 8–16 weeks, older puppy 4–6 months, adolescent 6–18 months, adult 18+ months, senior 7+ years, rescue). Stage determines session length, sessions per day, and default module priorities.
3. Goal-to-module mapping — Primary goal (from 12 options) maps to a defined set of cue training modules and issue management modules from the cue registry and issue module data files.
4. Breed modifiers — If a breed is specified, breed traits (energy, barking tendency, heat sensitivity, alone-time tolerance) adjust session intensity notes and caution flags but do not change the fundamental cue sequence.
5. Schedule generation — Daily session count and duration are set by life-stage rules and user-specified training time per day.
6. Plan assembly — Weekly roadmap, cue cards (steps, success criteria, progression/regression rules), management rules, reward strategy, and troubleshooting are assembled from the data files and returned as structured output.

All cue progressions follow positive reinforcement methods only. The engine enforces a no-aversives constraint: e-collar, prong, choke chain, alpha roll, and flooding approaches are never recommended. If aversive tool use is disclosed, a caution flag and humane alternatives note are added. Training plan outputs are educational starting points; they do not replace a qualified, certified professional trainer for complex behavioral issues.

Dog Food Calculator & Label Checker

The Dog Food Calculator estimates daily calorie needs and portion sizes using a standard veterinary metabolic formula:

• Resting Energy Requirement (RER): 70 × (body weight in kg)^0.75. This formula is consistent with WSAVA and Merck veterinary nutritional guidelines.
• Maintenance Energy Requirement (MER): RER multiplied by a life-stage and activity factor (e.g., 1.6 for an active adult, 3.0 for a young puppy, 1.2 for a spayed/neutered adult at rest).
• Size-class modifier: Toy and small breeds receive a slight upward metabolic adjustment; large and giant breeds receive a slight downward adjustment, reflecting known metabolic rate differences by size.
• Body Condition Score (BCS) adjustment: BCS is assessed on a 1–9 scale. Dogs below ideal (BCS 1–4) or above ideal (BCS 6–9) receive a portion adjustment toward an appropriate calorie target. The tool does not prescribe weight-loss rates above 1–2% of body weight per week.

Portion outputs are provided in dual units (cups/ml and grams/oz). The label assessment module evaluates the primary protein source and first five ingredients against AAFCO guidance. Safety gates check for urgent conditions (collapse, severe distress) and vet-first conditions (extreme underweight, extreme overweight, health conditions requiring dietary management). Dog Food Calculator outputs are educational guidance only; they do not replace veterinary nutritional assessment, especially for dogs with health conditions.

Dog Weight & Body Condition Checker

The Dog Weight & Body Condition Checker estimates ideal weight, daily energy needs, and weight-loss timelines using the same veterinary RER formula as the Dog Food Calculator:

• BCS-based target weight: Each BCS unit above 5 represents approximately 10% excess body weight (WSAVA Body Condition Score Chart). Target weight = current weight ÷ (1 + max(0, BCS − 5) × 0.10).
• RER/MER: Resting energy requirement = 70 × weight^0.75. Maintenance multipliers are drawn from AAHA 2021 Nutrition Guidelines and the Merck Veterinary Manual energy requirement table.
• Safe weight-loss rate: 1–2% of body weight per week, consistent with AAHA Weight Management Guidelines for Dogs and Cats.
• Muscle Condition Score (MCS): Assessed on a 3-point scale (normal/reduced/severely-reduced) per WSAVA MCS guidelines. Severe muscle loss triggers a mandatory vet-first gate.
• Safety gates: Urgent gate for acute illness symptoms. Vet-first gates for BCS 8–9 (obesity), significant muscle wasting, pregnancy/nursing, and known medical conditions. None of these conditions are managed by dietary adjustment alone in this tool.

This tool does not prescribe specific food brands, medications, or surgical interventions. Outputs are educational starting points for a conversation with your veterinarian.

Dog Crate & Bed Size Calculator

The Dog Crate & Bed Size Calculator produces home crate sizing, IATA airline pre-check estimates, bed recommendations, and an audit of the user's current crate. All calculations use interior (usable) dimensions, not exterior product dimensions.

• Home crate sizing: Uses the four-measurement A/B/C/D method recommended by the AKC and VCA: interior length = body length (A) × 1.1; interior height = standing height (D) × 1.1; interior width = shoulder width (C) × 1.25. Calculated dimensions are mapped to standard crate size bands (XS through XXL).
• IATA CR-1 pre-check: Airline cargo travel uses the IATA Live Animals Regulations Container Requirement 1 formula: L = A + B/2; W = 2 × C; H = D. This is a pre-check estimate only — each airline sets its own rules and must be contacted directly before booking.
• Brachycephalic adjustment: For flat-faced breeds (e.g., French Bulldog, English Bulldog, Pug), all IATA dimensions are multiplied by 1.10 per USDA APHIS guidance, to account for increased ventilation requirements and body proportion differences.
• Bed sizing: Flat bed length = body length × 0.8 + 4-inch buffer; flat bed width = shoulder width × 1.2. Bolster beds add 4–6 inches. Orthopedic style is recommended for dogs aged 8+ years or with mobility concerns, consistent with VCA guidance on orthopedic support.
• Puppy growth: For puppies under 12 months, the calculator recommends using a crate divider set to current size and re-measuring every 4–6 weeks as the puppy grows.
• Safety gates: Urgent gate for acute injury (blocks full plan). Vet-first gate for post-surgery/medical recovery. Caution gates for panic/escape history, chewing hazard, and collar-in-crate risk. Info gates for puppy growth and multi-dog households.

This tool does not guarantee crate safety, flight approval, or escape prevention. Outputs are sizing guidance only. Always verify interior dimensions with the specific product before purchase, and confirm airline requirements directly with your carrier.

Dog Name Finder

The Dog Name Finder is a browse and filter tool across 767 curated dog names. Filters operate as independent boolean conditions: names must satisfy all active filters simultaneously to appear. Available filters are gender (male/female/unisex), style (classic/trendy/cute/tough/nature), inspiration (mythology/nature/gaming/pop culture/food/sports), and length (short 1–4 chars/medium 5–7/long 8+). Names marked with a "Popular" badge are in the top 31 by historical usage frequency in the source dataset. Favorites are saved to browser localStorage and persist between visits.

Puppy First-Year Planner

The Puppy First-Year Planner builds a calendar-month timeline (month 1 = 0–4 weeks old) covering vet and vaccination milestones, socialization targets, training focus, supplies, watch-fors, and one-time costs, drawn from AAHA, AVSAB, AKC, and ASPCA guidance.

• Current-month placement: The puppy's age in weeks is mapped onto the 12-month timeline using the same age-range boundaries as the timeline data, so the plan opens on the month matching the puppy's current age and marks earlier months as complete.
• Training stages: Consecutive timeline months are grouped by training stage (young puppy, older puppy, adolescent) and paired with the same age-stage-rules.json priorities and cautions used by the Dog Training Plan Generator, so the two tools stay consistent.
• Recurring monthly cost: Calculated from a national-average annual cost by size class (toy/small/medium/large/giant), then scaled by the same lifestyle-tier (Budget/Typical/Premium) and US state cost-of-living multipliers used by the Dog Cost Calculator.
• One-time costs: Setup, veterinary, and training cost ranges come from the timeline data. Veterinary-category costs — and the optional spay/neuter line item — are adjusted by the state cost-of-living multiplier; setup and training costs are treated as nationally consistent retail prices and are not state-adjusted.
• Spay/neuter: Selecting "planned this year" adds a size- and state-adjusted surgery cost range to month 6, the typical decision-point month. "Already done," "not planned," and "not sure yet" add no cost.

All figures are general planning estimates, not quotes or guarantees. Vaccination schedules, socialization pacing, spay/neuter timing, and individual costs vary by region, breed, and veterinary recommendation — always confirm your puppy's specific schedule with your vet.

Dog Grooming Schedule Generator

The Dog Grooming Schedule Generator is a deterministic rule engine — there's no AI or photo analysis. It classifies a dog's coat into one of six archetypes, then derives a brushing, bathing, nail, ear, and dental schedule, a seasonal shedding note, and a home-vs-professional grooming recommendation, drawn from AKC, VCA, and AVMA guidance.

• Coat archetype classification: Six archetypes — short single, short double, long double, long silky, curly/wavy, and wire/rough — match the six categories used in the at-home grooming guide. If a breed is selected, its archetype comes from a 70-breed coat-type registry (high confidence). With no breed, users can pick a coat type manually (medium confidence); with neither, the tool defaults to a short double coat and flags the missing data (low confidence).
• Brushing, bathing, ears: Each archetype has a baseline brushing frequency and recommended tools, a bathing frequency range, and an ear-check cadence (weekly for the two long-coated archetypes, every two weeks for the rest), drawn from AKC brushing/grooming guides and VCA bathing/ear-cleaning guidance.
• Nail trims and dental care: Universal across all coat types — nail trims every 3-4 weeks (or whenever nails touch the floor when the dog stands naturally), per AKC guidance, and daily enzymatic-toothpaste brushing as the single most effective home dental step, per AVMA guidance, with periodic professional cleanings as dogs age.
• Shedding season: For double-coated archetypes (short double, long double, wire/rough), brushing steps up to its "shedding season" level during spring (Mar-May) and fall (Sep-Nov) to manage the heavier seasonal "coat blow." Single-coated archetypes get no seasonal note.
• Home-vs-professional routing: Archetypes that don't normally require clipping or stripping (short single, short double, long double) get a "home" recommendation. Archetypes that do (long silky, curly/wavy, wire/rough) get a "home + professional" recommendation with a typical interval — or, if the user prefers to avoid professional visits, a "home-only" recommendation with a trade-off note about the extra time and risk involved. Reported matting always overrides the above with a "see a professional" recommendation, regardless of coat type.
• Safety gates: Eight gates check for matting, skin issues, ear issues, overgrown nails, a dog that resists handling, senior/mobility needs, puppy age (under 12 months), and brachycephalic breeds. Matting and resistant-handling gates are caution-level; skin and ear issues are vet-referral gates; senior/mobility, puppy, and brachycephalic gates add adapted-routine guidance.

Outputs are general grooming guidelines based on coat type, not a veterinary or professional grooming assessment. Persistent skin or ear problems, severe matting, or a dog that's distressed by handling should be checked by a vet or professional groomer.

Dog Insurance Cost Estimator

The Dog Insurance Cost Estimator is a deterministic calculation built from industry-published rate patterns — there's no live quoting or insurer integration. It combines a breed-specific (or size-based) base premium with multipliers for age, location, coverage level, deductible, and reimbursement rate to produce a typical monthly figure and a realistic low-high range.

• Breed base premium and risk category: A 27-breed rate table, derived from NAPHIA industry data, assigns each breed a base monthly premium and a risk category (Low, Standard, Medium-High, High, or Very High) reflecting that breed's typical claims history. Breeds not in the table fall back to a size-based national average (toy, small, medium, large, or giant).
• Age factor: Premiums are lowest for ages 1-2 (factor 1.00) and rise steadily with age — roughly 1.7x by age 7 and roughly 3.5x by age 14+. Puppies under 1 year carry a 1.77x factor, reflecting enrollment-period pricing.
• Location adjustment: The same state cost-of-care index used in the Dog Cost Calculator adjusts the estimate up or down by state — for example, California and Hawaii carry some of the highest multipliers.
• Coverage, deductible, and reimbursement: Accident-only coverage costs roughly 42% of an accident & illness plan, while adding a wellness rider costs roughly 32% more. A $100 deductible costs about 28% more than the $500 baseline; a $1,000 deductible costs about 18% less. A 90% reimbursement rate costs about 16% more than 80%; 70% costs about 13% less.
• Range band: The final "typical" monthly figure is shown alongside a low-high range (roughly -15% to +20%), since real insurer quotes vary further by exact ZIP code, individual underwriting, and breed-specific risk adjustments this estimator doesn't model.
• Disclaimers: Every result is labeled an estimate for planning purposes, not a quote, and includes a reminder that pre-existing conditions are excluded, that a veterinarian visit doesn't replace insurer underwriting, and that comparing quotes from multiple insurers is the only way to get an exact price.

This is an estimate for planning purposes, not a quote. Actual premiums depend on the specific insurer, your dog's individual health history, and your exact ZIP code — always compare quotes from multiple insurers before choosing a policy.

Dog Food Safety Checker

The Dog Food Safety Checker is a database lookup tool — there's no AI or image recognition. It matches your search against a curated registry of 30+ human foods and ingredients, each rated against published veterinary toxicology sources.

• Ingredient database: Each entry includes a safety rating (safe, caution, dangerous, toxic), the specific toxic compound where applicable, published toxic and lethal dose thresholds per kilogram of body weight, symptom lists, onset timing, and source-backed emergency action guidance. Safe foods include serving preparation, nutrition highlights, and portion limits.
• Safety ratings: Four tiers — Safe (no known toxicity, fine with normal portions), Caution (conditionally safe with specific preparation, portion, or label checks), Dangerous (physical harm risk such as choking, splintering, or obstruction), and Toxic (contains compounds that are poisonous to dogs, with severity sub-levels of low, moderate, high, or critical).
• Weight-based dose calculator: For toxic foods with published dose thresholds, entering your dog's weight produces personalised toxic and lethal dose estimates. These are approximate thresholds from published veterinary literature — individual sensitivity varies and some toxins (e.g. grapes) have no established safe dose.
• Sources: All ratings are derived from ASPCA Animal Poison Control, AKC nutrition guidance, Merck Veterinary Manual toxicology sections, Pet Poison Helpline, and VCA Animal Hospitals. Each ingredient entry lists its specific source references.

This tool is for general guidance — it does not replace professional veterinary advice. If your dog has eaten something potentially harmful, contact your vet or an emergency animal poison helpline immediately.

Dog Behavior Checker

The Dog Behavior Checker is a search-and-triage tool — no AI diagnosis or automated assessment. It matches your description of a behaviour against a curated registry of 40 common dog behaviours, each written against published veterinary behaviour resources.

• Behaviour database: 40 entries covering vocalization, anxiety, aggression, destructive, compulsive, physical, neurological, and other categories. Each entry includes an urgency level, possible causes, common triggers, a "why this happens" explanation, step-by-step guidance, and clear "when to see a vet" criteria.
• Urgency triage: Four levels — Emergency (seek immediate veterinary care, e.g. head pressing, sudden disorientation), Vet-soon (schedule an appointment, e.g. aggression, excessive licking), Monitor (try the home steps and watch for changes, e.g. excessive barking), and Normal (common and usually harmless, e.g. zoomies, grass eating).
• Search matching: Exact name matching, alias matching (over 300 natural-language aliases like "won't stop barking" or "can't be left alone"), and partial matching — sorted by match quality and urgency (emergency results surface first).
• Sources: All content is derived from ASPCA behaviour resources, American College of Veterinary Behaviorists (ACVB), Merck Veterinary Manual, AKC training guidance, AVSAB position statements, and VCA Animal Hospitals. Each behaviour entry lists its specific source references.
• Reward-based only: No punishment-based, dominance-theory, or aversive training methods are included anywhere in the tool. All guidance follows the AVSAB position statement on humane dog training.

This tool is for educational guidance — it does not diagnose conditions or replace professional veterinary or behaviourist assessment. If your dog's behaviour is causing harm or has changed suddenly, contact your vet.

Breed Groups

The breed groups layer organises the 62 PupWiki breeds (and 207 AKC registered breeds) by their historical working purpose using the American Kennel Club's 7-group taxonomy, with the FCI's 10-group system used as a supporting lens.

What the data represents

Group membership describes historical selection pressure — the instincts and physical traits breeders selected for over many generations. Trait profiles (energy, trainability, handler focus, etc.) are group-level averages on a 1–6 scale. They describe a central tendency, not a property of every individual in the group. Individual dogs vary substantially within any group based on genetics, upbringing, socialisation, and environment.

Sources

Group membership data is sourced from the AKC official breed list and group structure (akc.org), verified against the FCI nomenclature (fci.be). Trait profiles are derived from the referenced breed standard characteristics and are reviewed against veterinary and behavioural references including the AVMA, WSAVA, and Merck Veterinary Manual.

What breed group data does NOT provide

• Individual dog behaviour predictions — group membership is not a guarantee of behaviour
• Medical or health advice — health concerns must be discussed with a qualified vet
• Training prescriptions — any training approach should be discussed with a qualified behaviourist or trainer
• Breed safety classifications — no breed or group is labelled categorically safe or unsafe

How group data is used across PupWiki

Group membership surfaces in breed profiles as contextual background, in quiz and comparison tools as one signal among many, and on the dedicated breed group pages as the primary subject. Group data never overrides breed-level facts, and tool results that reference group membership are explicitly framed as group-level tendencies, not individual predictions.

AI-assisted content workflow

PupWiki uses AI language models to assist with content drafting, formatting, and structural editing. This is disclosed because transparency about production methods matters for content trust.

What AI assistance does in our workflow:

• Drafts initial content structure from authoritative source material.
• Formats and restructures existing factual content.
• Generates template variations from validated data (e.g., breed profile sections from structured breed data files).

What AI assistance does not do in our workflow:

• Produce health, safety, or nutrition claims without editorial verification against Tier 1 sources.
• Generate breed scores, tool outputs, or cost figures — those come from deterministic code.
• Replace editorial review: all health and safety content is reviewed against authoritative sources before publication.

Tool outputs (Breed Fit scores, cost estimates, training plans, food fit results) are produced entirely by deterministic rule-based code, not language models. No AI inference influences scoring weights, safety gates, or tool verdicts.

Source authority tiers

Data freshness and update cadence

Different content types are reviewed on different schedules:

• Cost model inputs — Reviewed annually. US national average prices are updated from published industry and veterinary association cost surveys.
• Breed trait data — Reviewed when authoritative registry standards update, or when significant new veterinary literature changes breed-specific health guidance.
• Health watchlist items — Reviewed when Merck Veterinary Manual or major veterinary bodies publish updated guidance for specific breeds.
• Training plan engine rules — Reviewed when AVSAB or AAHA update their position statements or behavioral guidelines.
• Food fit formula constants — Reviewed when WSAVA nutritional guidelines update.
• Source registry — All source URLs are checked at least annually. Broken or outdated sources are flagged and replaced.

Each breed profile and guide page shows a "last reviewed" date. If content has not been reviewed recently and there is reason to believe it may be outdated, we include a note.

Data sources

Methodology last updated: 2026-06-06