SQL Exploratory Data Analysis (EDA) — SQL Server

Exploratory SQL project built on top of a Medallion data warehouse (Bronze → Silver → Gold) in SQL Server.
Focus: quick, trustworthy insights with small, targeted T-SQL scripts.

---

Introduction

This project performs exploratory data analysis (EDA) directly against a SQL Server warehouse.
It uses standardized Silver tables and business-ready Gold views to answer practical questions about customers, products, dates, and sales behavior.

Stack: SQL Server • VS Code (SQL Server extension) • Git Bash • GitHub

---

Scenario / Problem Statement

Analysts need to explore data without exporting to spreadsheets or Python/R.
However, the source systems are fragmented (CRM and ERP), and data quality varies.
This EDA project demonstrates how to:

• Explore structure and health of the database quickly
• Understand dimensions and date coverage
• Inspect core measures and distributions
• Run simple rankings and spot checks entirely in SQL

---

Datasets Used

The underlying datasets are provided by the vendored Data Warehouse:

CRM (source_crm/)

• cust_info.csv — customer master
• prd_info.csv — product master
• sales_details.csv — sales order lines

ERP (source_erp/)

• CUST_AZ12.csv — customer demographics
• LOC_A101.csv — customer country
• PX_CAT_G1V2.csv — product category metadata

These are available in this repo via https://github.com/Emadeddin-Beshtawi/sql-data-warehouse-project

---

Objectives

• Run EDA queries directly in SQL over cleaned data
• Use Gold views and Silver tables for reliable results
• Keep analysis modular: each script targets a specific theme
• Showcase SQL-native exploration (no external notebooks needed)

ETL Skill Highlight (inherited from DW)

• The warehouse performs BULK INSERT (Extract/Load) to Bronze
• Cleansing & standardization in Silver (dates, gender/marital/country, product lineage, de-dupe)
• Business-ready Gold (dimensions/fact) enabling robust EDA without re-cleaning

---

Repository Structure

sql-exploratory-data-analysis-project/
├─ scripts/
│  ├─ 01_database_exploration.sql
│  ├─ 02_dimensions_exploration.sql
│  ├─ 03_date_range_exploration.sql
│  ├─ 04_measures_exploration.sql
│  ├─ 05_magnitude_analysis.sql
│  └─ 06_ranking_analysis.sql
├─ deps/
│  └─ data-warehouse/                # vendored DW (tag v1.0.1)
├─ .gitignore
├─ LICENSE
└─ README.md

---

Conclusion

This EDA repo demonstrates SQL-only analysis on a clean warehouse surface:

• Start fast with database/dimension/date exploration

• Validate measures and distributions

• Produce quick top-N rankings and sanity checks

The approach scales because ETL work is centralized in the DW; analysts focus on questions, not cleanup.

---

Related Projects

• SQL Data Warehouse (Medallion): https://github.com/Emadeddin-Beshtawi/sql-data-warehouse-project

• SQL Data Analytics: https://github.com/Emadeddin-Beshtawi/sql-data-analytics-project