UCLA

A myriad of digital services mediate our daily lives. As a result, we are continuously generating digital traces, referred to as small data, that can be used to chronicle, characterize, and influence our behaviors and preferences. This thesis is focused on the notion of small data systems --- the services and tools designed for individuals to more directly and personally leverage their collective small data. We develop novel algorithms, toolsets and the system infrastructure to address cross-cutting challenges in small data systems. We evaluate the efficacy and feasibility of our contributions with real-world datasets, system deployments, and user studies in two application contexts: (1) Lifestreams, a toolset to facilitate the exploration of small data for personal behavioral analysis and chronic disease prevention, and (2) Immersive Recommendations, a new recommendation paradigm using small data to enable effective personalization for online services across the web.