The social sciences have long employed simulations and statistical models in their research methodology as testing and validation tools. Recent advances in complex networks have produced a set of new graph-based tools with which to study and characterize larges-scale, complex social data. We propose a swarm simulation platform that allows quick prototyping of agent-based social models, using an iterative simulation development cycle. This platform is designed to provide meaningful graph-based metrics of the social ties formed. A description of the system is given, as well as experimental results obtained from sample models.