Building Features from Nominal and Numeric Data in Microsoft Azure
Raw data rarely wins competitions or production models—feature quality does. This course teaches you how to engineer meaningful features from categorical and continuous data in Azure, cutting through the noise that derails real-world ML projects.
AIU.ac Verdict: Essential for data engineers and ML practitioners building pipelines in Azure who need to move beyond basic data loading. You’ll gain hands-on techniques for handling messy, mixed-type datasets. Note: assumes foundational Azure and data manipulation knowledge.
What This Course Covers
You’ll work through practical feature engineering strategies for nominal (categorical) data—encoding, grouping, and handling cardinality—alongside numeric transformations like scaling, binning, and outlier treatment. The course emphasises Azure-native tools and workflows, showing how to implement these techniques within Azure ML and data services rather than isolated notebooks.
Expect real scenarios: imbalanced categories, missing values, feature interactions, and performance trade-offs. Mike West structures this around building features that actually improve model accuracy, not just theoretical best practices. You’ll leave with a repeatable workflow for diagnosing weak features and strengthening them before model training.
Who Is This Course For?
Ideal for:
- Data Engineers: Building production ML pipelines in Azure who need to standardise feature preparation across teams.
- ML Engineers & Data Scientists: Transitioning from notebooks to Azure-based workflows and wanting to systematise feature engineering.
- Analytics Engineers: Responsible for data quality and transformation who want to bridge analytics and ML feature requirements.
May not suit:
- Complete Azure Beginners: You’ll need basic familiarity with Azure services; this isn’t an Azure 101 course.
- Advanced Feature Engineering Specialists: If you’re already shipping production models with sophisticated feature stores, the depth may feel introductory.
Frequently Asked Questions
How long does Building Features from Nominal and Numeric Data in Microsoft Azure take?
1 hour 19 minutes of video content. Plan 2–3 hours total if you’re following along with hands-on labs in Azure.
Do I need Azure credits or a subscription to complete this course?
Pluralsight typically provides sandbox environments for labs, but confirm access with your AIU.ac account. Some exercises may benefit from a personal Azure free tier account.
What’s the difference between this and general feature engineering courses?
This course is Azure-specific, teaching you how to implement feature engineering within Azure ML, Data Factory, and related services—not generic Python or R techniques.
Will this course cover machine learning model training?
No—the focus is strictly on preparing and engineering features. You’ll learn to build inputs that *feed* models, not train them.
Course by Mike West on Pluralsight. Duration: 1h 19m. Last verified by AIU.ac: March 2026.
