Ecommerce storefront
A year of orders across five joined tables, with best sellers, heavy buyers, a real Q4 peak and totals that reconcile to the cent.
An online retailer's 2025: 2,000 customers in ten countries, a 300-product catalogue, orders, line items and reviews. Every order total is exactly the sum of its own line items, so the joins and the arithmetic both hold when you check them. Around that it behaves like a shop: a few business customers place a third of the orders, cheap best sellers fill most baskets, consumers order from their phones at the weekend, reviews come from people who bought the product after it arrived, and November and December rise because there are more orders, not bigger ones.
- ecommerce
- retail
- star schema
- seasonality
Explore
Every table, profiled
Each column's type, spread, empties and most common values, measured from the CSVs in the download. Switch to the first rows to see the data exactly as it sits in the file.
order_items.csv
31,031 rows · 6 columns · 2 foreign keys
Keys
Every join resolves
5 foreign keys, 83,305 references checked against the table each one points at. None points at a row that does not exist.
order_items
- order_idorders.order_id31,031 rows checked0 orphans
- product_idproducts.product_id31,031 rows checked0 orphans
orders
- customer_idcustomers.customer_id11,081 rows checked0 orphans
reviews
- product_idproducts.product_id5,081 rows checked0 orphans
- customer_idcustomers.customer_id5,081 rows checked0 orphans
Checks
What we checked
- 0 orphaned foreign keys across all 5 relationships
- 0 orders dated before their customer signed up
- order_total equals the sum of its line items, exactly, for every order
- 0 products priced at or below cost (margins run 22% to 64%)
- 0 reviews by a customer who had no delivered order of that product
- Ratings are J-shaped (57% five-star, 6% one-star), and 36 of 226 well-reviewed products average under 3.5
- Top 10% of customers place 35.6% of orders; the top 20% of products are on 64% of lines
- November and December average 2.1x a January–September month while order value stays $386–456
- 63% of prices end in .99, and no city name appears in two countries
The same checks ship in the zip as INTEGRITY.txt, so you can rerun them in any SQL engine. How we verify
Questions, answered
Worth asking it
Each answer is the query's result on these exact files. Open one for its SQL and rows, then check your own query against it.
1Which category carries the best margin, and is it the one selling most?Apparel has the best margin (56.0%); Sports & Outdoors sells the most units.
-- Orders that were kept: cancelled and returned orders left out. select p.category, sum(oi.quantity) as units, round(sum(oi.line_total), 2) as revenue, round(100 * (1 - sum(oi.quantity * p.unit_cost) / sum(oi.line_total)), 1) as margin_pct from order_items oi join orders o on o.order_id = oi.order_id join products p on p.product_id = oi.product_id where o.status not in ('cancelled', 'returned') group by p.category order by margin_pct desccategory units revenue margin_pct Apparel 7,481 474,392.85 56 Beauty 10,994 321,028.53 54.7 Sports & Outdoors 17,489 729,716.81 42 Home & Kitchen 15,383 1,411,409.08 41 Electronics 11,206 1,146,088.1 29.7 2How much of revenue comes from the top 10% of customers, and which segment are they?The top 10% of buying customers bring 40.3% of revenue; most are Enterprise.
with spend as ( select customer_id, sum(order_total) as revenue from orders where status not in ('cancelled', 'returned') group by customer_id ), ranked as ( select customer_id, revenue, ntile(10) over (order by revenue desc) as decile from spend ) select c.segment, count(*) as customers, round(sum(r.revenue), 2) as revenue, round(100 * sum(r.revenue) / (select sum(revenue) from spend), 1) as share_of_all_revenue_pct from ranked r join customers c on c.customer_id = r.customer_id where r.decile = 1 group by c.segment order by revenue descsegment customers revenue share_of_all_revenue_pct Enterprise 112 1,094,204.41 26.8 Small Business 51 370,479.46 9.1 Consumer 30 177,933.04 4.4 3Which products are rated worst, and do they still sell?The worst rated, Aurea Hydrating Face Cream, Travel Size at 1.75 stars, still sold 114 units (percentile 54 of products).
with rated as ( select product_id, count(*) as reviews, avg(rating) as rating from reviews group by product_id having count(*) >= 5 ), sold as ( select product_id, units, round(100 * percent_rank() over (order by units), 0) as sales_percentile from ( select oi.product_id, sum(oi.quantity) as units from order_items oi join orders o on o.order_id = oi.order_id where o.status <> 'cancelled' group by oi.product_id ) s ) select p.product_name, p.category, r.reviews, round(r.rating, 2) as avg_rating, s.units as units_sold, s.sales_percentile from rated r join products p on p.product_id = r.product_id join sold s on s.product_id = r.product_id order by r.rating, r.reviews desc limit 10product_name category reviews avg_rating units_sold sales_percentile Aurea Hydrating Face Cream, Travel Size Beauty 8 1.75 114 54 Cedar Storage Containers Home & Kitchen 9 2.22 121 56 Verdant Hair Dryer Beauty 15 2.4 154 65 Trailhead Swim Shorts, Black Apparel 5 2.4 39 7 Kestrel Fitness Tracker, Pro Electronics 40 2.55 564 94 Stride Camping Tent, Pro Sports & Outdoors 24 2.58 391 90 Ashford Knit Beanie Apparel 13 2.62 127 58 Nimbus Portable Charger Electronics 32 2.63 387 89 Silkroot Vitamin C Serum, Travel Size Beauty 11 2.64 164 68 Voltix Wireless Earbuds Electronics 36 2.69 430 91 4What does the Q4 lift look like split by channel?A Q4 month against a January to September one: marketplace 1.98x the orders, web 1.82x the orders, mobile_app 1.76x the orders.
with m as ( select channel, extract(month from order_date) as month, count(*) as orders, sum(order_total) as revenue from orders where status <> 'cancelled' group by channel, extract(month from order_date) ) select channel, round(avg(case when month >= 10 then orders end) / avg(case when month <= 9 then orders end), 2) as q4_order_lift, round(avg(case when month >= 10 then revenue end) / avg(case when month <= 9 then revenue end), 2) as q4_revenue_lift, round(sum(case when month <= 9 then revenue end) / sum(case when month <= 9 then orders end), 2) as jan_sep_avg_order, round(sum(case when month >= 10 then revenue end) / sum(case when month >= 10 then orders end), 2) as q4_avg_order from m group by channel order by q4_revenue_lift descchannel q4_order_lift q4_revenue_lift jan_sep_avg_order q4_avg_order marketplace 1.98 2.12 186.49 199.06 web 1.82 1.83 497.72 501.19 mobile_app 1.76 1.7 367.36 356.15
Every query runs on the files in the download as they are. ANSWERS.sql holds all of them; run it in DuckDB from the dataset's folder.
Use it
Take it, or make one shaped like yours
Rebuild it byte for byte
The zip includes schema.yaml. Run misata generate --config schema.yaml with seed 20260721 and you get the same rows.
Your own tables, your own numbers
Describe the data you need in Studio and get connected tables with the totals you state, profiled like this and checked before you download.
Premium datasets
All premiumCard spending each month, debit and credit
2,500 US households at five banks and seven card issuers, two years
Five banks' raw statement lines, with every label behind them
Answer key
- description
- merchant_name
- category
- +6 more
- Rows
- 7,627,186
- Tables
- 14
- Checks
- 50 of 50
Free previewProfile and preview- transactions, rows per monthMay 2025 – Oct 2026
AML Transaction Monitoring
$5launch priceOne UAE bank, 10,000 customers, eighteen months
A UAE bank's whole trail from KYC to STR, with the laundering nobody caught
Answer key
- is_suspicious
- typology
- ring_id
- +7 more
- Rows
- 11,711,393
- Tables
- 16
- Checks
- 129 of 129
Free previewProfile and preview Weekly sales, at regular price and on promotion
Grocery chain, 10 stores, fiscal 2024 and 2025
Two fiscal years of a grocery chain, from the shelf to the receipt
Answer key
- baseline_units
- price_effect
- promo_lift
- +7 more
- Rows
- 2,790,852
- Tables
- 14
- Checks
- 117 of 117
Free preview18-question projectProfile and preview
Related free datasets
All 9- Retail star schemaA dimensional model you can actually practise joins on: one fact table, four dimensions, 47,051 rows, every join resolving and every total known.47,051 rows · 5 tables
- B2B SaaS subscription analyticsAccounts, seats, MRR, churn and support load, where company size actually drives the plan.42,240 rows · 5 tables
- Credit risk portfolio3,000 borrowers and 4,200 commercial loans, where a credit rating actually predicts whether the loan defaults, at the rate S&P's own data says it should.7,200 rows · 2 tables
- Network intrusion / netflow500 hosts and 20,000 labeled flows, where a port scan, a DDoS flood, and a brute-force attempt each carry the real statistical signature their MITRE ATT&CK technique actually leaves in traffic.20,500 rows · 2 tables
- Contact center queueing (Erlang C)4 queues, 2,400 half-hour intervals over a full business month, where agents_staffed is what Erlang C (1917) says is actually required to clear each queue's own SLA target, not a plausible-looking number.2,404 rows · 2 tables
- Clinical trial: sites, patients, visits and adverse eventsTwo trials across nine sites: enrolment, scheduled visits, dosing, lab biomarkers and graded adverse events, all in the right order.22,496 rows · 7 tables
Fully synthetic. No real person, company or transaction is represented, and no production data was read to make it. CC0: use it anywhere, no attribution needed.

