A Stochastic Block Model to model Sardex community currency
Sebastien Appleby, Trinity College Dublin.
06th of November 2025
Abstract:
Sardex is a pioneering community currency established in Sardinia 15 years ago to promote local businesses. This presentation explores the structural and dynamic properties of the network through the study of a 10-year dataset. The study investigates how network models can help understand and quantify economic dynamics between members. In particular, we focus on the velocity of exchanges. This metric, inspired by monetary economics, captures how rapidly value circulates within the network.
To model these patterns, we use the Stochastic Block Model (SBM) as we compare intra- and inter-cluster flows, trying to asses if the SBM effectively captures the structure of the Sardex transactions. Beyond the static analysis, we have identified potential extensions using the Dynamic SBMs to account for temporal evolution in group memberships and Graph Neural Networks (GNNs)—such as GCN, GraphSAGE, and GAT—to enhance predictive capabilities for community detection or transaction forecasting.