Open data studies on how people actually dress.
Plain-language, evidence-backed reports on AI fashion analysis, scored outfit datasets, and the behavioral patterns behind everyday dressing.
OutfitScore Research is the publishing arm of the OutfitScore platform. Every report draws on aggregated, de-identified data from real outfit submissions and is released under CC BY 4.0 for free reuse by journalists, researchers, and fashion writers. Methodology is documented in full. Datasets are available on request for academic and editorial use — write to [email protected].
The Anatomy of a Score
What 492 AI-analyzed outfits reveal about how people actually dress.
A data-driven study of 492 outfit submissions scored by AI between April and May 2026. Identifies the five dominant deficiency clusters — fit, accessories, texture, layering, and footwear — that account for nearly 90% of all improvement recommendations. Quantifies the 25-point gap between occasion-specific and undirected dressing, and proposes a ranked hierarchy of interventions by impact-per-effort.
The Psychology of Fashion Perfection Seekers
What 492 self-submitted outfits reveal about the pursuit of sartorial approval.
A behavioral analysis of the same 492-outfit corpus. Identifies six interlocking findings — the Pursuit Paradox, the Dunning-Kruger Wardrobe, the Comfort Zone Plateau, the Variance Paradox, the Asymmetric Self-Critic, and the Mirror Confessional — and proposes a perfectionism-vs-satisficing framework for escaping the score plateau most dressers find themselves stuck in.
The Anatomy of a Makeup Score
What 917 real faces reveal about modern beauty — and the one thing the industry refuses to talk about.
A data-heavy audit of 917 makeup analyses identifying the five deficiencies that explain almost all score loss. The headline finding: application technique, not product choice, is implicated in 28.6% of all recommendations — more than any single product category. Top-tier looks share three traits the beauty industry rarely foregrounds: even skin tone, cohesive palette, well-groomed brows. Includes a recommendation hierarchy and references to Etcoff, Mulhern, Tolentino, and Bobbi Brown.
The Clean Girl Verdict
Why 60% of real makeup submissions now embrace the no-makeup aesthetic — and what the data says about whether they're right.
A behavioral audit of the same 917-look makeup corpus. Roughly six in ten submissions describe themselves in some variant of natural / clean girl / no-makeup makeup terms; only eleven describe themselves as full glam. The AI rubric agrees: natural-leaning looks outscore glam-leaning ones by ~12 points. But the data also exposes the price of the aesthetic — the underlying skin, time, money, and racial archetype it quietly assumes. Includes a cultural cycle diagram and references to Bourdieu, Tolentino, Etcoff, and Vogue Business trend reporting.
About OutfitScore Research
Reports are authored by Saad, founder of OutfitScore, drawing on the aggregated analysis corpus of the platform. New issues are published approximately quarterly. All charts, figures, and findings may be reproduced with attribution. The underlying scoring methodology — a five-dimension rubric applied via multimodal large language models — is documented in Report No. 01, Section 1.
For media inquiries, citation guidance, or to request the underlying dataset for academic or editorial purposes, contact [email protected].