R for AEC Data Analysis
Preface
AEC projects generate rich, varied data, layer structures, 3D geometry, point clouds, sensor streams, and most of it is locked inside proprietary formats that R cannot easily reach. R for AEC Data Analysis is a practical guide to changing that.
This book uses the AutoDeskR R package (Govan 2024) and the AutoDesk Platform Services (APS) cloud API as the primary bridge to that data, translating design files, extracting geometry, connecting live sensor streams, and embedding interactive 3D models in Shiny dashboards. AutoDesk is the dominant platform in the AEC industry, so it’s a practical place to start. But the analytical techniques here, mesh analysis, layer analytics, digital twin patterns, MCP tool exposure, apply to any BIM or CAD data, wherever it comes from.
Whether you’re a data scientist who just inherited a folder of DWG files, or a BIM manager who wants to automate something that currently takes three software packages and a prayer, this book is for you.
License & Availability
This book is available online for free at paulgovan.github.io/AutoDeskR-Book and is released under a Creative Commons Attribution 4.0 International (CC BY 4.0) license — you are free to share and adapt the material as long as you give appropriate credit.
A print edition is available on Amazon if you prefer a hard copy.
What You’ll Learn
Each chapter covers a different piece of the BIM and CAD analytics ecosystem:
- Authentication — grab an OAuth token in one line and never think about it again
- Data Management — upload files to the cloud, wrangle buckets, and pull objects back down
- Model Derivative — translate design files into OBJ, STL, and SVF; extract geometry and metadata
- Design Automation — run DWG-to-PDF conversion in the cloud, no AutoCAD required
- Reality Capture — turn a set of overlapping photos into a 3D mesh
- Viewer — embed a live 3D model viewer in a Shiny app with two lines of code
- 3D Geometry & Meshes — import, measure, and visualise translated mesh files
- DWG & DXF Analytics — parse layer structure, extract attributes, compare drawing revisions
- Digital Twins — link live sensor data to BIM elements via AutoDesk Tandem
- MCP Server — expose AutoDeskR as an AI agent tool so LLMs can query BIM models directly
How This Book is Structured
The book is organised into eight parts. You don’t have to read them in order, each chapter is reasonably self-contained, but the parts do build on each other if you want the full picture.
| Part | Chapters | What you’ll do |
|---|---|---|
| Front Matter | Motivation, Acknowledgements | Understand the why |
| Getting Started | Getting Started | Install the package and make your first API call |
| AutoDesk Platform Services | Authentication → Viewer | The six main APS APIs, start to finish |
| 3D Geometry & Meshes | Reading → Visualisation → Point Clouds → Comparison | Import, measure, and compare translated mesh files |
| DWG & DXF Analytics | Layer Analysis → Attributes → Comparison → Reports | Extract meaning from drawing data |
| Digital Twins | Tandem Overview → Sensor Streams → Live Dashboards | Join live sensor readings to BIM elements |
| AI Integration | MCP Server | Expose AutoDeskR as a tool for AI agents |
| Reference | Case Study, Function Reference, Troubleshooting | Worked example, lookup tables, and help |
Prerequisites
To follow along you’ll need:
- R (version 4.0 or later) and, optionally, RStudio
- A free AutoDesk account with an APS application registered at aps.autodesk.com. The Getting Started chapter walks you through it step by step
- An internet connection (all the action happens in AutoDesk’s cloud)
No prior CAD or BIM experience assumed. If you’re comfortable running an R script, you’re ready.
Installation
Install the stable release from CRAN:
install.packages("AutoDeskR")Or grab the development version from GitHub if you like living on the edge:
devtools::install_github("paulgovan/AutoDeskR")How to Use This Book
New to APS? Start with Getting Started and work through the chapters in order, each one builds on the previous. Already know the APS basics? Jump straight to whatever chapter covers your use case. The Troubleshooting chapter is always there when something goes sideways.
All code examples are copy-paste ready and work with #| eval: false so they won’t execute during book rendering but will run just fine in your R session.
About the outputs in this book. Most examples call live AutoDesk Platform Services, which need your own credentials, uploaded files, and (for Digital Twins) a Tandem subscription, so they can’t run during book rendering. Where it helps to see the shape of a result, the book shows a representative output generated without the live API: tables, charts, and 3D views built from illustrative data. These are flagged in place with a Representative result note. Anything you must swap in to run the code yourself, bucket names, client IDs, file URNs, is called out with a Replace before running note.