### 🧹 Challenge 02 — About

**From Structure to Trust**

The grocery chain now has a solid database — but real-world data is rarely perfect.
In this stage, you’ll learn how to **detect, repair, and document** common data issues so the models you build later rest on reliable foundations.

After importing the schema from Challenge 01, you’ll run a complete **data-quality audit**: finding missing values, duplicates, unit mix-ups, and timestamp drift. Then you’ll apply **principled imputation techniques** to fix them while keeping the data’s meaning intact.

By the end, you’ll have a **clean, validated dataset** — one you can trust for forecasting, experimentation, and ROI modeling in future challenges.

---

#### 🧭 Series Context

Challenge 02 builds directly on the schema created in Challenge 01 (Spaghetti → Schema).
Here, you move from *organized data* to *trustworthy data* — learning the professional habits that separate working pipelines from fragile ones:

* documenting every change,
* checking assumptions, and
* proving reproducibility.

These same practices form the bridge to **Challenge 03 (Data → Signal)**, where trustworthy data becomes the foundation for time-series forecasting.

---

#### 🎯 Learning Goals

* Identify and describe common data-quality issues.
* Apply multiple imputation methods and justify your choices.
* Validate that cleaning hasn’t distorted key patterns.
* Create transparent documentation others can audit or extend.

---

#### 💡 Why It Matters

Every great model depends on clean, believable data.
Challenge 02 helps you build that trust — turning “structured but messy” datasets into solid ground for analytics and machine learning.