The increasing availability of smartphones, cameras, and wearables with instant data sharing capabilities, and the exploitation of social networks for information broadcast, heralds a future of real-time information overload. With the growing excess of worldwide streaming data, such as images, geotags, text annotations, and sensory measurements, an increasingly common service will become one of data summarization. The objective of such a service will be to obtain a representative sampling of large data streams at a configurable granularity, in real-time, for subsequent consumption by a range of data-centric applications. This paper describes a general-purpose self-summarizing storage service, called Social Trove, for social sensing applications. The service summarizes data streams from human sources, or sensors in their possession, by hierarchically clustering received information in accordance with an application-specific distance metric. It then serves a sampling of produced clusters at a configurable granularity in response to application queries. While Social Trove is a general service, we illustrate its functionality and evaluate it in the specific context of workloads collected from Twitter. Results show that Social Trove supports a high query throughput, while maintaining a low access latency to the produced real-time application-specific data summaries. As a specific application case-study, we implement a fact-finding service on top of Social Trove.