In this project, I aim to represent search processes in "correlation spaces", carried out by users of the data exploration software Mirador (http://fathom.info/mirador/). These spaces are metric, that is, are provided with a notion of distance, which allows to calculate the separation between random variables according to their degree of statistical correlation. The databases used in these visualizations are the Global Development Indicators(World Bank), and the Behavioral Risk Factors (Center for Disease Control, USA). I will also describe the application of these and other tools in the derivation of predictive models for clinical care.