Implementations directed to identifying variables affecting a cost of healthcare include actions of receiving data from a plurality of data sources, the data relating to cost of providing healthcare, providing a data model based on the data, processing the data model using non-parametric analysis to provide a non-parametric result, the non-parametric result including a first and second variable, automatically processing the data model to correlate at least the first variable of the non-parametric result to the second variable of the non-parametric result to provide a correlation result, processing the correlation result using parametric analysis to provide a parametric result including at least one variable that affects the cost more than one or more other variables of a plurality of variables of the data model, and providing at least one data visualization for display, the at least one data visualization providing at least one graphical representation of the parametric result.