Energy storage can help integrate local renewable generation, however the best deployment level for storage remains an open question. Using a data-driven approach, this paper simulates 15-minute electricity consumption for households and groups them into local communities of neighbors using real locations and the road network in Cambridge, MA. We then simulate PV for these households and use this framework to study battery economics in a high PV adoption, high electricity cost scenario, in order to demonstrate significant storage adoption. We compare the results of storage adoption at the level of individual households to storage adoption on the community level using the aggregated community demands. Under the simulated conditions, we find that the optimum storage at the community level was 65% of that at the level of individual households and each kWh of community battery installed was 64-94% more effective at reducing exports from the community to the wider network. Therefore, given the current increasing rates of residential battery deployment, our research highlights the need for energy policy to develop market mechanisms which facilitate the deployment of community storage.