The sample dataset
Everything behind this site, in the formats you would receive in production. Snapshot 2026-09-01, postcode 10001.
All CSVs, the raw keyword JSON, the data dictionary and the README in one archive.
Individual files
| File | Rows | Provenance | Contents | |
|---|---|---|---|---|
observed_snapshot_real.csv | 1,560 | Observed | Every ranked listing across 12 sub-categories on the snapshot date — rank, price, rating, reviews, ad status. | Download |
keyword_rankings_raw.json | 2,827 | Observed | Organic and paid rank for 21 head keywords, three result pages deep, including the full cat-tree set. | Download |
client_catalog.csv | 36 | Observed | The complete SONGMICS HOME + VASAGLE catalogue as listed on Wayfair US. | Download |
analysis_sku_namespace.csv | 12 | Observed | SKU-prefix fingerprint analysis — SONGMICS-namespace product ranking under Wayfair private labels. | Download |
analysis_rank_band.csv | 5 | Observed | Cross-section by rank band: reviews, rating, price, brand mix, visibility gradient. | Download |
analysis_share_of_shelf.csv | 12 | Observed | Per-category brand composition, ad density and the top-12 review entry bar. | Download |
dim_product.csv | 1,560 | Observed | SKU dimension table. | Download |
dim_category.csv | 12 | Observed | Category dimension table with Wayfair category IDs. | Download |
fact_daily_rank.csv | 27,912 | Illustrative panel | 28-day daily panel. Final day is the real snapshot; preceding days are back-simulated to show the feed's schema and shape. | Download |
fact_weekly_snapshot.csv | 3,904 | Illustrative panel | Weekly roll-ups with rank volatility, review deltas and estimated units. | Download |
manifest.json | — | Observed | Run metadata, calibration constants and generator seed. | Download |
DATA_DICTIONARY.md | — | Observed | Every column, typed and defined, with join keys. | Download |
README.md | — | Observed | Collection method, cadence options and history availability. | Download |
On provenance
What is real
Collected from live Wayfair pages on the snapshot date
The 1,560 category listings, the 2,827 keyword results, the 36-SKU client catalogue, the SKU-namespace analysis and all four analysis files. Every finding on this site and in the accompanying report is drawn from these files only.
What is illustrative
Clearly labelled, seeded and reproducible
The 28-day daily panel and the weekly roll-ups built on it. We cannot hand you history we have not collected yet, and we will not pretend otherwise — so the panel is generated from the real snapshot using a rank→visibility curve calibrated on the observed cross-section. Every such row carries is_observed = False. Use it to validate your pipeline and dashboards, not to draw market conclusions.