Free star schema generator
Generate any data warehouse schema.
Any domain. Any scale.
Pick a domain, see the schema canvas, generate a preview with verified referential integrity, and download it. Every foreign key resolves. Every *_key column links cleanly. No SQL Server to install, no Kaggle login, no fixed dataset to work around.
Step 1: Choose a domain
Dimensions included
Built for
Step 2: Inspect the schema
Classic retail star schema: fact_sales surrounded by date, product, customer, store, and promotion dimensions. Foreign keys drawn as edges. Click any table or column to inspect properties.
Step 3: Generate and verify
How Misata compares
Every alternative is a fixed dataset.
This one shapes itself to you.
| Tool | Row scale | Domains | Formats | Custom targets | FK proof | No-signup CSV |
|---|---|---|---|---|---|---|
Misata DW GeneratorThis page | Custom (2K–10M+) | Any domain | CSV, SQL, Parquet | Preview free | ||
AdventureWorksDW | Fixed ~10M | MS SQL DW only | SQL backup restore | - | - | SQL install required |
Kaggle Retail Star Schema | Fixed ~300K | Retail only | CSV | - | - | Kaggle login required |
Kaggle DAX Practice | Fixed ~100K | Retail only | CSV | - | - | Kaggle login required |
Star Schema Benchmark (SSB) | Scalable but complex | Lineorder only | DB-specific setup | - | - | DB setup required |
How the generation works
01
Declare, don't sample
Unlike SDV, Gretel, or MOSTLY AI, Misata never trains on real data. You state what the schema should look like and what the numbers should add up to. The engine solves for rows that satisfy those constraints.
02
Referential integrity guaranteed
Every foreign key in every dim table is a valid primary key in the referenced table. Zero orphaned rows. The proof is attached to every run and checked with DuckDB, independent of the generator.
03
Power BI, SQL, Parquet ready
Files are named fact_* and dim_*. Column types are consistent. Numeric columns use tabular-style formatting. Load directly into Power BI, Tableau, DuckDB, or any SQL database without transformation.
Frequently asked questions
Need a dataset shaped to your exact numbers?
Declare exact revenue curves, churn rates, or conversion targets. The engine generates data where the declared aggregate holds to the cent, not approximately.