Code

Open-source agent-based models, each with the paper that documents it.

Every model I have published is open source. Most come with a paper that documents the design and the calibration. Everything lives on github.com/BAFurtado.

PolicySpace3

The current platform. Heterogeneous families, firms, local governments and a financial institution interacting across labour, goods, housing, rental and mortgage markets, in spatially detailed Brazilian urban concentration areas (ACPs/CONURBs). Built for prospective analysis: carbon taxes, PlanHab housing scenarios, commuting.

Code

PolicySpace2

A spatial agent-based model of Brazilian metropolitan regions with an explicit real estate market, mortgage credit, firms, labour and municipal taxes. Used to compare tax and housing policy alternatives.

Code JASSS paper Book (PT) arXiv

VIDA — domestic violence

A simulation of domestic violence dynamics under social distancing, built to support policy responses to violence against women in Brazil.

Code JASSS paper

eco_cars

A regional model of green car adoption with taxes and subsidies, following van der Vooren and Brouillat (2015), reimplemented in Python.

Code arXiv

PolicySpace

The first generation of the model: municipal tax distribution across metropolitan regions.

Code arXiv

ML simulates ABM

Machine learning surrogates trained to emulate an agent-based model, so that large parameter sweeps become affordable.

Data & code arXiv

Axtell’s firms on networks

Endogenous dynamics of firms made up of networked agents — an exploratory replication and extension.

Code Documentation (PT)

A simple spatial model of the economy

Tools for policy: the minimal spatial agent-based economy the later models grew out of.

CoMSES JASSS paper