US retail product data · Verified 2026-09-18 to 2026-10-18
Clean, barcode-verified product and price data from 8 US retailers.
1,173,683 verified US retail SKUs (968,718 Big 3 Flagship, 824,552 unique GS1 GTINs) from Target, Walmart, Amazon US (Buy Box), Meijer, Kroger, Fry's, Aldi and Publix, matched across stores by exact GTIN parity. No fuzzy matching. Delivered as Parquet, SQLite or CSV.
200 rows free, no sign-up · Starter pilot from $250 · Big 3 Flagship $4,950 · Pay by card, ACH or invoice
core_gold_v2_verified_products.parquet · 5 of 1,173,683 rows
Retailer
Product
GTIN
Price
GS1 check digit valid
Every row passes the check
1,173,683
Verified catalog SKUs
686,093 noisy records quarantined from 1,859,776 raw · 685,341 strict_1p
824,552
Unique GS1-valid GTINs
282,864 leading-zero SKUs preserved as zero-padded text
420,234
Cross-store pairwise matches
Across 272,445 multi-store entities · 9.9% Big 3 / 10.51% All-8 clean parity
100%
GS1 Modulo-10 pass rate
968,718 Big 3 SKUs (100% direct_storefront)
Coverage & provenance
What's in the dataset
Eight US retailers with explicit seller_type and price_origin provenance tags. Big 3 Flagship delivers 968,718 100% direct_storefront SKUs; Core Gold v2 adds regional banners and transparent cross-referenced grocery pricing.
Target448,675 SKUs
288,098 strict_1p + 160,577 marketplace_3p · 448,675 direct_storefront · from 735,311 raw
Walmart266,867 SKUs
192,278 strict_1p + 74,589 marketplace_3p · 266,867 direct_storefront · from 423,373 raw
Amazon US (Buy Box)253,176 SKUs
253,176 amazon_buy_box · 253,176 direct_storefront · from 284,882 raw
Meijer120,402 SKUs
120,402 strict_1p · 120,402 direct_storefront · from 143,669 raw
Kroger52,014 SKUs
52,014 strict_1p · price_origin: upc_cross_reference (0 direct_storefront) · from 124,242 raw
Fry's Food Stores32,338 SKUs
32,338 strict_1p · price_origin: upc_cross_reference (0 direct_storefront) · from 93,061 raw
Aldi171 SKUs
171 strict_1p · 171 direct_storefront · from 1,309 raw
Publix40 SKUs
40 strict_1p · 40 direct_storefront · from 53,929 raw (strict QA quarantine)
1,859,776 raw records cleaned down to 1,173,683 verified SKUs (686,093 quarantined). Kroger (52,014) and Fry's (32,338) carry upc_cross_reference price provenance while Target, Walmart, Amazon US (Buy Box), Meijer, Aldi, and Publix carry direct_storefront shelf prices.
Every UPC-A and GTIN-14 passes the official GS1 Modulo-10 check-digit test before entering the catalog. All 824,552 unique GTINs pass, and 282,864 leading-zero SKUs stay formatted as clean strings.
Explicit seller & price provenance
686,093 corrupted or unverified raw rows were quarantined out of 1,859,776 raw inputs. Remaining rows tag exact seller_type (685,341 strict_1p, 253,176 amazon_buy_box, 235,166 marketplace_3p) and isolate multipack outliers separately from clean unit parity.
Ready to query in seconds
Delivered as Apache Parquet (~117.6 MiB All-8 / ~100.6 MiB Big 3), SQLite, or CSV. Query all 1,173,683 verified rows directly in DuckDB, Polars, PyArrow, Databricks, or BigQuery.
Free sample
Explore 200 real rows
Search and filter 200 verified records across all 8 retailers (including seller_type and price_origin provenance). Commercial storefront URLs and cross-store match tables ship in paid licenses.
Six ways a team can use the data. Each one includes the SQL to try it and what you would receive. These are illustrations, not customer stories.
Pricing
National beverage and snack price gap audit
How a pricing team can replace brittle HTML scrapers with exact GTIN joins across Walmart, Target, Amazon US (Buy Box), and Kroger while filtering out multipack outliers.
260,813 clean parity pairs · 10.51% mean spread (420,234 total pairs) · under 40 ms DuckDB join
SELECT a.upc_gtin, a.brand, a.title, a.price_usd AS walmart_usd, b.price_usd AS target_usd,
ROUND(ABS(b.price_usd - a.price_usd) / LEAST(a.price_usd, b.price_usd) * 100, 2) AS spread_pct
FROM 'big3_verified_products.parquet' a
JOIN 'big3_verified_products.parquet' b ON a.upc_gtin = b.upc_gtin
WHERE a.retailer = 'Walmart' AND b.retailer = 'Target'
AND a.is_multipack = 0 AND b.is_multipack = 0
ORDER BY spread_pct DESC LIMIT 100;
What you receive: Verified Parquet partition + SHA-256 digest (eb0d440fab85508034afa2973c5fab41353c03144c3f450f33cbfc26bfa7f74e) separating 232,252 clean parity pairs (9.9% Big 3 mean spread) from 17,794 multipack outliers.
AI and ML
Ground-truth product graph for LLMs
How an ML team can avoid variant mix-ups (single unit vs. case pack) in retail models. Modulo-10 GTIN keys identify the exact SKU, and multipack flags isolate bulk bundles.
SELECT upc_gtin, ARRAY_AGG(DISTINCT retailer) AS banners, MIN(price_usd) AS floor_usd, MAX(price_usd) AS ceiling_usd
FROM 'core_gold_v2_verified_products.parquet'
WHERE is_multipack = 0
GROUP BY upc_gtin HAVING COUNT(DISTINCT retailer) >= 3
ORDER BY (MAX(price_usd) - MIN(price_usd)) DESC;
What you receive: Normalized Parquet & SQLite index with provenance columns (seller_type, price_origin) and GS1 check-digit proof for every row.
Investment research
Multi-retailer grocery and household price tracker
How a research team can track identical national-brand baskets across Target, Walmart, Amazon US (Buy Box), Meijer, Kroger, Fry's, Aldi and Publix for inflation nowcasting.
SELECT category, retailer, COUNT(*) AS basket_skus, ROUND(AVG(price_usd), 2) AS basket_avg_usd
FROM 'core_gold_v2_verified_products.parquet'
WHERE seller_type = 'strict_1p' AND price_origin = 'direct_storefront'
GROUP BY 1, 2 ORDER BY 1, 4 ASC;
What you receive: Weekly Snappy-compressed Parquet delta feed delivered via pre-signed S3 URI, GCS, or Snowflake Secure Share.
Category management
Store brand vs. national brand pricing
How a category manager can isolate private-label GS1 company prefixes and compare them with national brand equivalents across grocery and household departments.
Good & Gather · Great Value · Simple Truth · Fred's & Aldi labels
SELECT gs1_prefix, brand, retailer, COUNT(*) AS skus, ROUND(AVG(price_usd), 2) AS mean_shelf_usd
FROM 'core_gold_v2_verified_products.parquet'
WHERE brand IN ('Good & Gather', 'Great Value', 'Simple Truth', 'Threshold')
GROUP BY 1, 2, 3 ORDER BY 4 DESC;
What you receive: GS1 prefix-mapped private-label vs. national brand taxonomy matrix in Parquet, SQLite, and CSV.
Regional pricing
Midwest and regional grocery matrix
How a team can compare regional grocers (Meijer direct storefront plus Kroger & Fry's UPC cross-reference baselines) with Walmart and Target using GTIN-14 parity.
SELECT retailer, price_origin, COUNT(*) AS skus, ROUND(AVG(price_usd), 2) AS mean_usd
FROM 'core_gold_v2_verified_products.parquet'
WHERE retailer IN ('Meijer', 'Kroger', 'Frys', 'Aldi', 'Publix')
GROUP BY 1, 2 ORDER BY skus DESC;
What you receive: Regional grocery table with explicit price_origin flags and exact GTIN-14 join keys.
Insurance
Replacement cost index for personal property
How an insurance team can support proof-of-loss schedules and homeowner property inventories using verified 1P and Buy Box retail prices.
824,552 verified barcodes · direct storefront & GS1 provenance
SELECT canonical_gtin14, brand, title, model, MIN(price_usd) AS floor_rcv_usd, MAX(price_usd) AS ceiling_rcv_usd
FROM 'core_gold_v2_verified_products.parquet'
WHERE price_origin = 'direct_storefront' AND is_multipack = 0
GROUP BY 1, 2, 3, 4;
What you receive: Actuarial-grade replacement cost lookup table keyed by UPC-A, GTIN-14, and manufacturer model.
Data dictionary
23 public fields, one consistent v2 schema
Every public sample record mirrors the verified v2 Parquet and SQLite table columns below (commercial storefront link fields are excluded from public downloads and delivered in paid tiers). Click SQL on any row for a DuckDB query.
Field
Type
Required
Example
Description
Example query
retailer
VARCHAR(32)
Required
"Target" | "Walmart" | "Amazon"
Canonical retailer name across the 8 US retailers.
DuckDB / Snowflake query for retailer
SELECT retailer, COUNT(*) AS skus FROM 'core_gold_v2_verified_products.parquet' GROUP BY 1 ORDER BY 2 DESC;
SELECT seller_type, COUNT(*) FROM 'core_gold_v2_verified_products.parquet' GROUP BY 1 ORDER BY 2 DESC;
price_origin
VARCHAR(32)
Required
"direct_storefront" | "upc_cross_reference"
Price capture origin: direct_storefront (1,089,331) or upc_cross_reference (84,352 for Kroger & Fry's).
DuckDB / Snowflake query for price_origin
SELECT retailer, price_origin, COUNT(*) FROM 'core_gold_v2_verified_products.parquet' GROUP BY 1, 2 ORDER BY 3 DESC;
last_verified_utc
TIMESTAMPTZ
Required
"2026-10-08T05:31:25.557182+00:00"
ISO-8601 UTC verification timestamp (2026-09-18 to 2026-10-18). Storefront URLs reserved for commercial tiers.
DuckDB / Snowflake query for last_verified_utc
SELECT MIN(last_verified_utc), MAX(last_verified_utc) FROM 'core_gold_v2_verified_products.parquet';
Verification
Check it yourself
Test any barcode with the same GS1 check-digit math we apply to every record, and compare your delivery files against published SHA-256 checksums.
Validate any barcode
The last digit of a UPC or GTIN is a check digit. It is calculated from the other digits using 3x and 1x positional weights (GS1 Modulo-10), so a mistyped or invented code fails the test.
Try a sample:
Enter a barcode or pick a sample to see the calculation.
Verified v2 delivery checksums (2026-09-18 to 2026-10-18)
Each licensed v2 snapshot ships with an immutable SHA-256 digest and exact byte size. Copy any digest to verify your local Parquet or SQLite build.
What changes when product matching is based on exact barcodes instead of scraped HTML, title similarity or closed vendor keys.
What you compare
This dataset
Raw web scraper APIs
Fuzzy-matching vendors
Legacy syndicated feeds
How products are matched
Exact GTIN barcode match, verified with the GS1 check digit
Unlinked raw pages, often missing UPCs
Title similarity that can confuse similar variants
Proprietary closed product keys
Provenance & QA quarantine
686,093 noisy rows quarantined; explicit seller_type & price_origin columns on all 1,173,683 rows
Unlabeled 1P vs. 3P shelf prices mixed together
Varies by vendor
First-party only, but opaque and delayed
Format and delivery
Parquet, SQLite and CSV, delivered right away with SHA-256 digests
Rate-limited JSON webhooks
Locked into a web app, with export caps
Legacy SFTP flat files and slow onboarding
Cost to start
$250 Starter Pilot, $2,450 Single-Retailer, $4,950 Big 3, or $8,950 All-8 Master (fits $10k P-Card)
Ongoing proxy and maintenance cost
Usually an annual contract
Usually a multi-year enterprise agreement
Pricing
Unified commercial licensing
Start with a 10k-SKU pilot slice, license the Big 3 direct-storefront flagship, or acquire the complete All-8 Core Gold v2 corpus under the $10,000 P-Card cap.
Starter Pilot Slice
Fast schema, join, and benchmark evaluation slice.
$250launch special · $500 standard
10,000 verified SKUs
250 clean UPC match pairs
Full provenance (seller_type, price_origin) & SHA-256 verification
Single-Retailer Deep Dive
Complete catalog for one banner or Amazon US 253k.
$2,450one-time
Target (448,675), Walmart (266,867), Amazon US (Buy Box) (253,176), or Meijer (120,402)
Need a custom retailer subset (e.g. Walmart + Amazon 520,043 SKUs or Target + Amazon 701,851 SKUs) or SaaS OEM terms? Request a custom quote.
Get the data
Request a quote or invoice
Send a request and we reply with an invoice or ACH details within 4 business hours.
FAQ
Questions
Why are storefront links and cross-store price tables omitted from the free sample?
The free 200-row sample includes all 23 verified catalog, provenance, and GS1 barcode fields, while leaving out commercial storefront URL columns and pre-computed cross-store arbitrage spread tables. Full storefront links and pairwise match tables ship with paid commercial tiers, starting with the $250 Starter Pilot Slice.
What provenance guarantees apply to Amazon, Kroger, Fry's, and Publix?
Every record explicitly states seller_type (strict_1p, amazon_buy_box, or marketplace_3p) and price_origin (direct_storefront or upc_cross_reference). Amazon US carries 253,176 amazon_buy_box records (direct_storefront), Kroger (52,014) and Fry's (32,338) carry upc_cross_reference pricing, and Publix retains 40 strict QA-verified rows from 53,929 raw captures.
How does pricing fit under a $10,000 corporate P-Card limit?
Both our Big 3 US Flagship ($4,950 one-time for 968,718 direct-storefront SKUs and 250,046 pairs) and our complete All-8 Master Core Gold v2 ($8,950 one-time for 1,173,683 SKUs and 420,234 pairs) sit cleanly under standard $10,000 P-Card caps so teams can procure via card invoice, ACH, wire, or Net-30 with W-9.
Who issues the invoice and W-9?
All commercial licenses, W-9 documents, and SHA-256 delivery manifests are issued by That One Guy Pondering LLC (Florida Division of Corporations Document #L26000523238).
Try the data before you buy.
Browse 200 real rows, check the barcodes yourself, then request a quote when you are ready.