Abstract
This paper employs dynamic graph analytics to examine the evolution of events in the lives of individuals participating in welfare support programs. By focusing on relational dynamics rather than transactional services, this approach captures the complexity of interactions and their impact over time. The representation of network dynamics aids in understanding the present impact of various welfare interventions, facilitates dialogue among case workers regarding the intensity and outcomes of interventions, and redefines accountability as an evolving process. Using administrative data from jobseeker support programs in northern Italy and a relational support program in the UK, we constructed dynamic graphs to track relationships between users and service providers. These graphs allowed us to extract features and apply the ODDNET anomaly detection algorithm to identify significant change points. Subsequent discussions with domain experts helped interpret these points. Our methodology underscores the importance of structural changes in relational networks, signaling the need for further investigation when significant shifts occur. This analysis provides insights into the effectiveness and sustainability of support programs, highlighting the need for ongoing relational engagement beyond program durations. The use of open-source tools and existing datasets ensures cost-effectiveness and replicability, while maintaining privacy and fairness by anonymizing data and focusing on roles and categories rather than personal information. This study demonstrates the potential of dynamic network analytics as a powerful tool for evaluating and improving relational services within welfare programs. .
Recommended Citation
Montaletti, Giampaolo, "e. Welfare interventions are akin to Multi-Organizational ones. Complexity of Relational, more than Transactional, Services requires Dynamic Graph Analytics to redefine Accountability" (2024). OISI Workshop 2024. 9.
https://aisel.aisnet.org/oisiworkshop2024/9