D5.9 Complex modelling and artificial populations for agent-based simulations

The goal of this deliverable is to provide an overview of the current implementation of the Synthetic Populations Generator (SPG), a service developed in the framework of the FOSSR project for the extraction of synthetic populations. These are a class of algorithms for the extraction of statistical properties of a population from aggregated or separated data. This is a critical challenge when micro-level data is not available due to privacy concerns or information being scattered over different datasets. Identify such micro-level information can benefit quantitative and simulation methods for the study of population dynamics such as agent-based modelling. The goal of SPG is to support researchers and policymakers with a friendly service to extract micro-level information from aggregated data. To this aim, the algorithm implements a general product rule of joint probabilities from marginal data identified from the input data and adjusting for joint counts provided by the user. The output of the service is a table containing the number of individuals for each intersection profile, iterative also by space units. Furthermore, users can download a script for agent-based modelling ready to be initialized with the same table extracted for the creation of the synthetic population. Aligning with the Open Science framework, the service is delivered through three channels. Open code developed in Python is accessible through a GitHub repository, a local demo is available, and the service also runs through a web-app in the Virtual Research Environment (VRE) developed by FOSSR-Lab. The three modalities align with the Open Science perspective of the FOSSR project and are envisaged to facilitate future integrations of the service in the FOSSR cloud in the future.