Fabric Retail Analytics

End-to-end retail analytics project built in one day with Microsoft Fabric, Dataflow Gen2, Lakehouse, SQL, PySpark, DAX and Power BI.

The project uses the Brazilian Olist e-commerce dataset to build an analytics workflow from raw data ingestion and transformation through data modeling to an interactive Power BI dashboard.

Architecture

The solution follows an end-to-end analytics workflow in Microsoft Fabric.

CSV → Dataflow Gen2 → Lakehouse → SQL & PySpark → Semantic Model → Power BI

Data Pipeline

Dataflow Gen2
Raw Olist CSV files are ingested and cleaned before being loaded into the Fabric Lakehouse.

Lakehouse
Cleaned data is stored as Delta tables and provides the source layer for downstream transformations.

SQL
SQL views create the analytical fact and dimension structures used by the semantic model. Separate queries are included for initial data exploration and data quality validation.

PySpark
A reusable date dimension is generated programmatically and persisted as a Delta table for time intelligence.

Semantic Model
The analytical model connects sales with product, customer, seller and date dimensions. DAX measures provide KPIs, year-over-year comparisons and dynamic formatting used in the report.

Dashboard

The Power BI report provides an overview of retail sales performance, including:

• Total Revenue
• Total Orders
• Average Order Value
• Items Sold
• Year-over-Year Performance
• Top Product Categories
• Top Customer States
• Category Performance by Revenue, Review Score and Items Sold

The category performance view combines commercial performance and customer feedback to identify high-performing and underperforming product categories.

Repository Structure

Fabric-Retail-Analytics/
│
├── docs/
│   ├── architecture.png
│   ├── dashboard.png
│   └── semantic-model.png
│
├── pyspark/
│   └── create_dim_date.py
│
├── sql/
│   ├── 01_data_exploration.sql
│   ├── 02_data_quality_checks.sql
│   ├── 03_order_reviews_view.sql
│   ├── 04_fact_sales_view.sql
│   ├── 05_dim_product_view.sql
│   ├── 06_dim_customer_view.sql
│   ├── 07_dim_seller_view.sql
│   └── 08_order_product_reviews_view.sql
│
└── README.md

Tech Stack

Microsoft Fabric · Dataflow Gen2 · Lakehouse · SQL · PySpark · Power BI · DAX

Dataset

This project uses the public Brazilian E-Commerce Public Dataset by Olist, containing information about orders, customers, products, sellers and customer reviews.mer reviews.