Our approach to data science reveals a new translational research paradigm. It starts at the point of translation through partnering stakeholders and focusing on their challenges. The result is a "research pull" versus a "research push". The pull happens through working many case studies in multiple domains and watching the synergies and overarching research needs emerge. A key component of our approach is to first focus on massive repurposing of existing data in the conceptual development work and let this drive our understanding of the data and method gaps that need to be filled. Data pipelines need to evolve all data – big, small, experimentally collected, and opportunistic – need to be discovered, wrangled, and repurposed into analyses.
Our research model includes conceptual development (the left-hand box) and the development of new approaches for key outcomes (the right-hand box). The conceptual development is composed of two parallel but interdependent tracks – the instantiation of a data pipeline in the context of a case study. The outcomes are to fill methodological and data gaps by developing new measures and validating and evaluating these measures.