Drop in a wide, repetitive spreadsheet. We mine the functional dependencies and propose normalized tables — dimensions plus a fact table — each downloadable as CSV.
Restructuring one wide table into several related tables so each fact is stored once. It removes redundancy (a customer's email repeated on every order row) and the update anomalies that come with it. The usual target is third normal form (3NF): every non-key column depends on the key, the whole key, and nothing but the key.
Column A determines column B (A → B) when each value of A always appears with the same value of B — like customer_id → email. Dependencies decide the split: columns determined by the same attribute belong in their own table.
No — it's mined from the data you loaded, so a dependency that happens to hold in this file may not hold universally (with more data, two customers could share a name). Treat it as a strong, data-driven first draft to review.
Imagine what we'd build for your team — data pipelines, dashboards, and infrastructure engineered to the same standard, on your data.
Talk to amanar.ai →