The project MAP-MO – Modena’s Marginalization Awareness Project proposes the creation of an innovative, data-driven and participatory knowledge infrastructure in the Modena socio-health district. Its objective is to generate continuous, high-quality information on the socio-economic conditions of adults facing severe marginalization - a population that is difficult to reach, highly heterogeneous, and deeply embedded in the local socio-economic context.
The project addresses a critical knowledge gap by combining cutting-edge AI technologies with a collaborative governance model. Unlike traditional observatories, MAP-MO is designed as a participatory system of knowledge production, where data are co-generated and validated by institutional actors (Municipality of Modena, AUSL Modena) and Third Sector organizations (outreach teams, shelters, support centers). This co-production ensures both the robustness of the data and their immediate relevance for policy and service design.
Building on the promising results of a 2024–2025 FAR Departmental pilot, which tested the feasibility of profiling the homeless population using administrative and qualitative data, MAP-MO will scale up by: i) deploying AI-driven data integration and cleaning tools, leveraging pre-trained Large Language Models (LLMs) and agent-based systems to merge heterogeneous sources and guarantee reliable datasets for predictive analytics; ii) developing predictive models of the spatial distribution, needs, and service trajectories of marginalized groups, enabling proactive and evidence-based interventions; iii) introducing AI agents as intelligent intermediaries, capable of autonomously extracting relevant information, activating the most appropriate machine learning models, and translating complex outputs into clear, interactive insights for decision makers and practitioners.
The expected impact is twofold. First, the project aims to advance the state of the art in the use of AI and LLM-based agents for social policy research, with strong potential for replication in other European urban contexts. Second, the project has the objective to strengthen local welfare governance by providing decision makers, health authorities, and Third Sector actors with a shared, adaptive, and actionable evidence base to address marginalization more effectively.
Thanks to the established commitment of key local stakeholders and the availability of diverse data sources, the project is highly feasible. At the same time, its methodological innovation and interdisciplinary design make it scalable and transferable well beyond the Modena case. Moreover, MAP-MO will serve as a scientific and training platform, fostering collaboration between researchers and practitioners and strengthening capacity building in data-driven welfare governance. In this way, the project positions itself as a pioneering model for integrating participatory approaches, AI, and predictive analytics in the governance of social vulnerabilities.