This course discusses advanced techniques for transforming, reducing, selecting, and discovering structure in data without relying on target labels. Learners gain hands-on experience with a range of unsupervised feature engineering methods, including Principal Component Analysis (PCA), Robust PCA, Singular Value Decomposition (SVD), Autoencoders, Variable Clustering, and Non-Negative Matrix Factorization (NMF). The course also covers visualization of high-dimensional data using t-SNE and demonstrates how to deploy and score unsupervised learning models in real-world analytical workflows. Through practical demonstrations and real-world use cases, participants learn how to extract meaningful information from complex datasets and thereby improving model efficiency, uncover latent patterns, and build topic discovery and recommendation solutions.

The course focuses on the data width aspect of analytics, that is, techniques related to variables and feature engineering. To explore the data depth aspect, which focuses on observations and pattern discovery, consider taking the companion course Unsupervised Machine Learning Series: Pattern Discovery Methods.