New Book Release: Practical Data Science Engineering
Building Reusable Workflows and Pipelines in Python
I am pleased to announce the publication of my new book:
Practical Data Science Engineering: Building Reusable Workflows and Pipelines in Python
Data science is often taught as a collection of techniques. One chapter covers data cleaning, another introduces visualization, and later sections explore machine learning algorithms.
What is frequently missing is a clear answer to a fundamental question:
How do these pieces fit together into a complete, reusable system?
This book was written to address that challenge.
Rather than focusing solely on individual tools or isolated notebook examples, the book approaches data science as an engineering discipline. Readers learn how to transform notebook-centric experimentation into reusable workflows that are modular, reproducible, maintainable, and scalable.
Throughout the book, readers build reusable components for:
- data loading and validation
- exploratory data analysis (EDA)
- missing-value handling
- outlier treatment
- feature scaling and normalization
- dataset splitting
- feature engineering
- model training and evaluation
- workflow orchestration
The book culminates in the construction of a complete reusable workflow architecture for tabular machine learning projects.
Included Companion Resources
Readers receive access to:
- chapter source code
- datasets and configuration files
- workflow templates
- downloadable resources
- the complete reusable workflow project
The final workflow system is also available as an independent repository that readers can clone, run, and extend for their own projects.
Who Should Read This Book?
This book is designed for:
- Python developers entering data science
- data analysts seeking stronger engineering practices
- machine learning practitioners building reusable workflows
- researchers working with reproducible pipelines
- anyone who wants to move beyond notebooks and think in systems
From Notebook to Workflow
The journey described throughout the book can be summarized simply:
Loading → Exploring → Cleaning → Splitting → Scaling → Engineering → Modelling → Evaluating → Assembling
The goal is not merely to build models.
The goal is to build workflows that others can run, understand, extend, and trust.
I hope this book helps readers develop the engineering mindset necessary to create practical and reusable data science systems.
Thank you for your support and interest.
— Shouke Wei, PhD
Deepsim Press
