A machine learning model is only as useful as the business question behind it. Jason Hodson builds this practical guide around three use cases, taking each from raw data to a deployed prediction.
Readical Key Points
- First edition, paperback, 440 pages, published 2026 by Rheinwerk Computing
- Three worked use cases, each built on a separate dataset: a retail case, customer retention and crime predictions
- Covers data exploration and cleaning, regression, decision trees, random forests, gradient boosting, clustering, validation metrics and interpretability
- Setup chapters use GitHub and Anaconda; sample code is available for download
Hodson moves through the full life of a model. Readers set up their tools, explore and clean the data, and choose an approach with a decision framework that favors the simplest model that works. Later chapters test results with validation metrics and k-fold cross-validation, explain what a model is doing, iterate on it, then put it into production, monitor it and measure its business impact.
The book serves analysts, data professionals and technically minded managers who need machine learning to answer a real business question. Because the three use cases return at every stage, a reader can follow one project from first look at the data to a monitored model. A closing chapter covers learning with generative AI. No book can supply a team's own data; what this one supplies is a repeatable path for working with it.
About the Author
Jason Hodson's career has centered on data roles for close to a decade, including forecasting work built on applied machine learning, and he has mentored colleagues in analytics and programming. He taught himself before completing a master's in business analytics.
Condition: Brand New and Readical Ready.
Readical™ is an independent bookstore and reading-products brand built around a simple idea: books matter most when what we learn leaves the page and enters our lives.
Apply the Page. Get Readical™.
