Animarium: an open, reproducible pipeline for synthetic populations of Italian cities, from ISTAT sources to open data (Tech Report v1)

A technical report introduces "Animarium," an open and reproducible pipeline for generating synthetic populations of Italian cities. This initiative created detailed synthetic data for eleven Italian municipalities, comprising over 1.8 million individuals and nearly 900,000 households, entirely from publicly available aggregate data sources. The methodology emphasizes transparency and verifiability, certifying each source with licensing and data affordances, and employing a multi-layered approach for attribute generation and placement.

Why it matters

This initiative provides a robust and transparent framework for generating synthetic population data from public aggregates. Such a capability can significantly enhance data-driven policy formulation, urban planning, and resource allocation without compromising individual privacy, fostering innovation in data utilization for public good.

Key insights

• Animarium generates synthetic populations for Italian cities using only publicly available aggregate data.
• The project successfully created synthetic data for eleven Italian municipalities, including 1,814,317 individuals and 887,937 households.
• Data sources include ISTAT census and register tables, census-section counts, the national civic-address register, public-use survey microdata, and municipal open-data portals.
• Each data source is certified with its license, fingerprint, and declared affordances, ensuring transparency and reproducibility.
• The generation pipeline uses a maximum-entropy joint model for demographic attributes and whole-vector donation for attitudinal and health variables.
• Individuals are placed to specific census sections, single years of age, and addresses, with households constrained by census size distribution.
• The pipeline demonstrates a method for creating rich synthetic datasets from disparate aggregate sources.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2608.27111