3D Data Science with Python

Building Accurate Digital Environments with 3D Point Cloud Workflows

Get the Book • About • Quick Start • Chapters • Resources • Author

---

Get the Book

Master the art of transforming raw 3D data into actionable intelligence. This comprehensive guide takes you from foundational concepts to cutting-edge techniques in 3D machine learning, deep learning, and spatial AI—all through hands-on Python implementations.

What You'll Build: Custom 3D viewers from scratch Analytical applications ML/DL pipelines for point clouds Automated reconstruction systems

---

About

Welcome to the official companion repository for "3D Data Science with Python" by Florent Poux, published by O'Reilly Media.

This repository provides all code examples, datasets, and supplementary resources to accelerate your learning journey.

Why This Book?

Feature	Description
Hands-On Learning	Every concept is reinforced with real-world datasets and production-ready Python code
Complete Coverage	From point cloud fundamentals to PointNet deep learning architectures
Build Real Tools	Create your own 3D viewers, analytical apps, and ML pipelines from scratch
Industry-Ready Skills	Prepare for careers in robotics, autonomous vehicles, geospatial analysis, digital twins, and more

A Note from the Author

"My journey into 3D data science began not with algorithms but with mud-caked boots as a land surveyor. One evening, staring at a .csv file of 2,587 data points, I had my spark moment: 'Couldn't this be automated?'

That question changed everything. This book is for you—the curious mind ready to unlock the potential of 3D data. Groundbreaking discoveries often begin with a simple question. Are you ready to ask yours?"

— Florent Poux

---

Quick Start

Prerequisites

• Python 3.8 or higher
• Windows, macOS, or Linux

Installation

pip install numpy pandas matplotlib open3d scikit-learn pyvista torch

For detailed environment setup instructions, refer to Chapter 3 of the book.

---

Chapters

Part	Chapter	Title	Code
I	1	Introduction to 3D Data Science	View
	2	Resources and Software Essentials	View
	3	3D Python and 3D Data Setup	View
II	4	3D Data Representation and Structuration	View
	5	Developing a Multimodal 3D Viewer with Python	View
	6	Point Cloud Data Engineering	View
	7	Building 3D Analytical Apps	View
III	8	3D Data Analysis	View
	9	3D Shape Recognition	View
	10	3D Modeling: Advanced Techniques	View
	11	3D Building Reconstruction from LiDAR Data	View
IV	12	3D Machine Learning: Clustering	View
	13	Graphs and Foundation Models for Unsupervised Segmentation	View
	14	Supervised 3D Machine Learning Fundamentals	View
V	15	3D Deep Learning with PyTorch	View
	16	PointNet for 3D Object Classification	View
	17	The 3D Data Science Workflow	View
	18	From 3D Generative AI to Spatial AI	View

---

Resources

Primary Learning Hub

All supporting materials—datasets, code, exercises, and complementary courses—are hosted at the 3D Geodata Academy.

Additional Links

Resource	Link
O'Reilly Book Page	oreilly.com/library/view/3d-data-science
3D Geodata Academy	learngeodata.eu
Official Errata	O'Reilly Errata Page

---

Citation

If you use the code or concepts from this book in your research, please cite:

@book{poux2025_3d_data_science,
  title     = {3D Data Science with Python: Building Accurate Digital Environments with 3D Point Cloud Workflows},
  author    = {Poux, Florent},
  year      = {2025},
  publisher = {O'Reilly Media},
  isbn      = {978-1098161309}
}

---

Author

Florent Poux, Ph.D. Florent is a leading authority in 3D data science, combining academic rigor with industry innovation. He serves as: Head Professor at the 3D Geodata Academy Researcher & Lecturer at top European universities Innovation Director for FrenchTech120 companies With over a decade of experience, Florent's expertise bridges deep academic research with practical 3D product development.

---

License

This repository uses a dual-license model:

Status	License	Commercial Use
Book Owner	Permissive (MIT-style)	Allowed
Non-Book Owner	CC BY-NC 4.0	Not Allowed

Own the book? You have full commercial rights to use, modify, and distribute the code.

Don't own the book yet? You can still use the code for learning and non-commercial purposes under CC BY-NC 4.0.

See the full LICENSE for details.

---

Ready to transform how you work with 3D data?

_{Built with passion for the 3D data science community}