Momentum One Academy · lab
Analyze Data with Apache Spark in Microsoft Fabric
This guided lab makes the Spark workflow in a Fabric Lakehouse concrete. Participants work with a small sales-order dataset, define a typed PySpark DataFrame, explore it with transformations and aggregations, and compare file-based and table-based persistence.
What the lab covers
- Creating a Lakehouse and preparing order data for Spark analysis.
- Defining a typed DataFrame schema instead of relying on inferred columns.
- Filtering, grouping and aggregating sales orders with PySpark.
- Writing and reading Parquet data partitioned by year and month.
- Persisting a Delta table and querying it with Spark SQL.
- Reviewing visual results and the practical effect of the work on a Spark session.
Learning outcome
By the end of the lab, participants can explain how a typed DataFrame moves from raw files to partitioned Parquet and a governed Delta table. They can identify when a partitioning choice supports efficient filtering and when Spark SQL offers a familiar access layer over Lakehouse data.
Workshop connection
This lab supports Module 04 — Data Cleaning & Transformation with Spark. It gives the team a shared foundation for discussing schemas, transformations, persistence formats and the hand-off between engineering-oriented Spark work and SQL-oriented analysis.
Access note
The detailed lab guide, screenshots and working materials are provided as part of the training experience. This page is an overview only; it does not publish the exercise instructions, code, participant assets or sample-data links.
Plan your workshop