Hands-On Healthcare Data: Taming the Complexity of Real-World Data

Healthcare analytics begins long before a model is trained. The records, claims, clinical vocabularies, and text behind the analysis must first be found, understood, and made to work together. In Hands-On Healthcare Data, Andrew Nguyen focuses on that demanding groundwork—and on how sound data preparation can make advanced analysis more meaningful.

Start with the realities of healthcare data

The book maps the varied sources that data teams encounter, including electronic health records, claims, clinical registries and trials, and digital health tools. Nguyen also considers how data collection methods shape what researchers and analysts can learn, distinguishing prospective from retrospective studies.

Connect systems, structures, and terminology

Readers are introduced to relational databases as well as property graphs, hypergraphs, and RDF databases. The discussion then turns to controlled vocabularies, terminologies, and ontologies, including CPT, ICD, LOINC, RxNorm, SNOMED CT, and the Unified Medical Language System. These topics help explain why bringing two datasets together often involves more than matching column names.

Follow the work through real-world examples

Case studies explore electronic health records, medication harmonization, publicly accessible datasets, and the combination of claims data with EHR information. The book also addresses extracting structured information from clinical text and the choices involved in normalizing medical data—useful context for readers assessing how tools and data models fit a particular problem.

From harmonized data to analytics

Later chapters cover feature engineering, graph-based machine learning, graph embeddings, federated learning and analytics, and clinical natural language processing. The closing chapters return to the broader task of merging datasets, bridging technical and business teams, and recognizing that data harmonization is not solely a technical exercise.

For readers working across data science and healthcare

This book is suited to data scientists, informatics practitioners, and technical readers who work with healthcare information or want to understand the domain’s distinctive data challenges. Its emphasis is practical: build a clearer picture of the data and its relationships before relying on advanced analytics to answer difficult questions.