Agent-based simulations can be a powerful tool for exploring possible evolutionary trajectories, community structures, or social network processes. Much of their power comes from the ability to run vast quantities of simulations and cover a broad parameter space. Here we present the SimDataCollection package, which automates running simulations over wide swaths of parameter space defined by all possible combinations of user-specified parameter values and gathering data from agent-based simulations built using the MASON Multi-agent Toolkit. The SimDataCollection package was employed in the two following simulations of the formation of aggregation preferences based on repeated interaction, or familiarity, in fission-fusion societies. A variety of species across taxa preferentially associate and cooperate with familiar individuals(those they have interacted with before) over those which they have not interacted with. This tendency to aggregate and cooperate with familiar individuals occurs even in fission-fusion societies, characterized by the frequent splitting and combining of groups, resulting in highly dynamics populations, which are highly prevalent in nature and should make the formation of particular partnerships difficult. In our first simulation, we employed an evolutionary agent-based model to illustrate that a preference for aggregating with familiar individuals could in fact make cooperation evolutionarily stable in such fission-fusion societies. Our second simulation focused on the effect of fission-fusion dynamics and the formation of associations over repeated interaction on the emergence of social networks, which we quantified using a variety of network approaches. These simulations present an initial view into the possible power of fission-fusion dynamics to effect social structure and therefore evolutionary outcomes, however, these simulation results can only go from conjecture to scientific truth when combined with empirical results.