Research — what Amazon review data can and cannot tell you
Independent data notes from the Amazon Reviews 2023 corpus: review-weighted ratings, star distributions, verified-purchase shares and yearly volume — with the limits stated on every page.
AMZ Reviews Data is an independent review-data research project. Our first beat is cookware: 18.1M Amazon reviews across 1,186,539 products, analyzed for defects, durability and long-term satisfaction. The cookware findings , answers and comparisons live in the main section. This desk publishes the underlying data notes for categories we have processed end to end.
What we publish here
- Data notes — one per category: review-weighted averages, star distribution, verified-purchase share, yearly volume, and an explicit “what this does not tell you”.
- Methodology — definitions, thresholds and the limits of a review snapshot.
- A comparison tool — look up two entities side by side. No scores, no rankings.
We do not publish a page per ASIN, brand, store or manufacturer. Those pages were retired in October 2026: they restated the same counts at scale and added nothing a reader could not get from the data note plus the comparison tool.
How to read a data note
Every figure is an aggregate. The average rating is review-weighted, not an average of product ratings, and 2023 is a partial year because the dataset ends in September 2023. When the sample cannot support a conclusion, we say so rather than fill the page.
Browse by category
Each data note covers one category: what the reviews show, and what they do not.
| Category | Reviews | Products | Reviews / product |
|---|---|---|---|
| Office Products | 12,841,907 | 709,449 | 18.1 |
| Baby Products | 5,930,003 | 204,671 | 29.0 |
| Amazon Fashion | 2,500,939 | 826,108 | 3.0 |
Tools
Sample size. This page is built from
Data vintage. Source: UCSD Amazon Reviews 2023 (data through September 2023). Historical statistics are not a statement about a product's current quality, availability or price. Review volume is not sales volume.
Scope & deduplication. One review per reviewer per product is counted; ratings without review text still count toward distributions. Coverage is uneven across years and categories, which is why the sample size is shown with every table. See the methodology for the exact definitions.
No raw reviews. We publish aggregated counts and word frequencies only — never full review text, reviewer identities or images.
Affiliate disclosure. Outbound links to Amazon are Amazon Associates links; if you buy through them we may earn a commission at no extra cost to you. Links point to a product page or to an Amazon search for the brand / store / category. We never display live prices or stock because those change constantly.