Skip to content

Geospatial Modeling with Python

Sergio Souza Costa — LambdaGEO, UFMA

Work in progress

This book is an active draft, not a finished text. Some material was assembled from existing course notes and translated into English with the assistance of AI tools; many chapters are still incomplete and will be revised with real-data examples as the underlying datasets and packages are finalized. The book is expected to reach a complete first edition by December 2027, as part of an ongoing research project on the DisSModel framework. Feedback and corrections are welcome via the book repository.

This book comes in two volumes, sharing one site and one numbering.

Volume I — Foundations (Ch 1–20)

A self-contained geographic data science and scientific Python course — no DisSModel required. Python fundamentals, Pandas, data cleaning, vector and raster geospatial analysis, spatial statistics, and the simulation paradigms (cellular automata, discrete-event simulation) built by hand before any framework enters the picture. If your interest is geospatial Python on its own, this volume is the whole book you need.

  • Part I — Scientific Python for Researchers Core tools and practices: Python fundamentals, Pandas, data cleaning, EDA, and software engineering for reproducible science.

  • Part II — Geographic Data Science A dual-substrate approach covering both vector and raster data models, multidimensional arrays, spatial relationships, and raster-vector integration.

  • Part III — Foundations of Spatial Simulation Cellular automata, discrete-event simulation with salabim, NumPy vectorization, and the performance problem that motivates DisSModel.

Volume II — The DisSModel Ecosystem (Ch 21–33)

Picks up exactly where Volume I leaves off and hands its concepts to DisSModel, the Python-native spatial modeling framework this book's own research group develops — installation, every simulation paradigm as a framework, domain case studies, infrastructure, and a TerraME/LUCCME migration guide. (The dissmodel package's own API reference lives with the code, at dissmodel.github.io/dissmodel; this volume is the narrative path to it, not a substitute.)

  • Part IV — DisSModel: Core and Paradigms Installing and building models with DisSModel, then each simulation paradigm in turn — system dynamics, cellular automata, agent-based modeling.

  • Part V — Domain Modeling: Land Use & Coastal Systems DisSLUCC's land-use change models and a full coastal dynamics case study.

  • Part VI — Data & Infrastructure Reproducibility, the DisSModel Platform, and spatial data cubes.

  • Part VII — Scaling, Migration & Reference Ensemble scenarios, migrating an existing TerraME/LUCCME model, and how to contribute to the ecosystem.

Part I — Scientific Python for Researchers

Chapters 1–5 are fully independent of DisSModel and can be read standalone.

Ch Title Notebook
1 The Scientific Python Ecosystem part1/ch01_ecosystem.ipynb
2 The Geospatial Python Toolbox part1/ch02_toolbox.ipynb
3 Tabular Data with Pandas part1/ch03_pandas.ipynb
4 Data Cleaning and Exploratory Analysis part1/ch04_cleaning_eda.ipynb
5 Software Engineering for Scientific Python part1/ch05_software_eng.ipynb

Part II — Geographic Data Science

A dual-substrate treatment of spatial data: vector and raster as complementary models. Chapters 6–16 are fully independent of DisSModel.

Ch Title Substrate Notebook
6 Introduction to Spatial Data Both part2/ch06_spatial_intro.ipynb
7 Vector Data with GeoPandas Vector part2/ch07_vector.ipynb
8 Raster Data with NumPy and rasterio Raster part2/ch08_raster.ipynb
9 Multidimensional Arrays with Xarray Raster part2/ch09_xarray.ipynb
10 Spatial Relationships and Weights Both part2/ch10_weights.ipynb
11 Point Pattern Analysis Vector part2/ch11_pointpatterns.ipynb
12 Exploratory Spatial Data Analysis Both part2/ch12_esda.ipynb
13 Spatial Regression Vector part2/ch13_regression.ipynb
14 Clustering and Regionalization Vector part2/ch14_clustering.ipynb
15 Visualizing Spatial Data Both part2/ch15_visualization.ipynb
16 Raster-Vector Integration Patterns Both part2/ch16_integration.ipynb

Part III — Foundations of Spatial Simulation

Simulation paradigms built by hand, before DisSModel is introduced. All four chapters are readable without any DisSModel knowledge.

Ch Title Notebook
17 Paradigms of Spatial Simulation part3/ch17_paradigms.ipynb
18 Cellular Automata from Scratch part3/ch18_ca.ipynb
19 Discrete-Event Simulation with salabim part3/ch19_des.ipynb
20 The Performance Problem — and the Solution part3/ch20_performance.ipynb

Part IV — DisSModel: Core and Paradigms

Where the framework itself takes over — installation, architecture, and each simulation paradigm from Part III revisited with DisSModel doing the bookkeeping.

Ch Title Notebook
21 Introducing DisSModel part4/ch21_dissmodel.ipynb
22 Building Models with DisSModel part4/ch22_building.ipynb
23 System Dynamics with DisSModel part4/ch23_sysdyn.ipynb
24 Cellular Automata with DisSModel part4/ch24_ca_dissmodel.ipynb
25 Agent-Based Modeling with DisSModel part4/ch25_abm.ipynb

Part V — Domain Modeling: Land Use & Coastal Systems

Ch Title Notebook
26 Land Use and Cover Change Modeling part5/ch26_lucc.ipynb
27 Case Study — Coastal Dynamics part5/ch27_coastal.ipynb

Part VI — Data & Infrastructure

Ch Title Notebook
28 Reproducibility and Experiment Provenance part6/ch28_provenance.ipynb
29 Running Models with the DisSModel Platform part6/ch29_platform.ipynb
30 Spatial Data Cubes part6/ch30_disscube.ipynb

Part VII — Scaling, Migration & Reference

Ch Title Notebook
31 Ensemble Scenarios and Sensitivity Analysis part7/ch31_ensemble.ipynb
32 Migrating from TerraME/LUCCME to DisSModel part7/ch32_migration.ipynb
33 Architecture and Contributing part7/ch33_architecture.ipynb

How to Use This Book

Each chapter is a Jupyter notebook. You can read it as a book or run it interactively. Code cells are self-contained within each chapter.

Parts I and II require no knowledge of DisSModel and are suitable for readers interested in geographic data science alone. Part III introduces simulation concepts independently before Part IV hands the same problems to the framework. Parts V through VII assume familiarity with DisSModel's core API from Part IV.

Installation

pip install geopandas rasterio xarray zarr libpysal salabim

For Part IV onward, which use DisSModel directly:

pip install dissmodel

Extension packages (dissmodel-ca, dissmodel-sysdyn, dissmodel-abm, disslucc-continuous, disslucc-discrete, brmangue-dissmodel) aren't on PyPI yet — install each one straight from GitHub, e.g.:

pip install "git+https://github.com/DisSModel/dissmodel-ca.git"

Source Code

All notebooks and supporting code are available at:

Citation

If you use this material in your research or teaching, please cite:

Costa, S. S. (2028). Geospatial Modeling with Python.
LambdaGEO Research Group, Federal University of Maranhão (UFMA).
https://lambdageo.github.io/geospatial-modeling-python

LambdaGEO Research Group · Federal University of Maranhão (UFMA) lambdageo.github.io