## 🧭 Overview — Challenge 02  Errors → Integrity

A grocery chain now runs on a clean relational database — but reality isn’t clean.
Missing transactions, drifted product codes, timestamp quirks, and duplicate records all erode confidence in analytics.
Your mission: **detect, repair, and validate** these flaws so every metric and model rests on data that can be trusted.

**Dr. V's Slide Decks :** [The Architecture of Validity](https://starshiptutor.substack.com/p/the-architecture-of-validity), [Challenge 01](https://starshiptutor.substack.com/p/challenge-01-spaghetti-schema-294), [Challenge 02](https://starshiptutor.substack.com/p/challenge-02-errors-integrity), [Challenge 03](https://starshiptutor.substack.com/p/challenge-03-data-signal)
**Meet the Avatars :** [Dr. Elena Vasquez](https://starshiptutor.substack.com/p/introducing-dr-elena-vasquez), [Evan Morales](https://starshiptutor.substack.com/p/introducing-evan-morales), [Maya Chen](https://starshiptutor.substack.com/p/introducing-maya-chen), [Daniel Reeve](https://starshiptutor.substack.com/p/introducing-daniel-reeve)
**Kaggle Pages :** [Challenge 01](https://www.kaggle.com/competitions/challenge-01-spaghetti-schema-2026), Challenge 02, [Challenge 03](https://www.kaggle.com/competitions/challenge-03-data-signal-2026)

This second challenge in the *Grocery Retail – Driven by Data* series transforms “structured” data into **reliable** data.
You’ll move from schema design to **systematic data auditing and principled imputation** — learning how transparency, documentation, and statistical rigor turn noise into signal without distortion.

By the end, you’ll deliver a reproducible data-quality workflow that others can run and verify: audits, repair logs, and post-imputation validations.
Integrity becomes the bridge between architecture and analysis — the moment when data science grows from *organized* to *trustworthy*.