Our presentation advocates for a cloud-based approach to analytics and insights computing-- it can be described as “data democratization”.  In our case this entails staging large datasets on the cloud and then using R to access and reshape the data. R can also be used to perform all the analytics, visualize the uncovered relationships and even dashboard the critical indicator variables.  We provide a detailed discussion of staging, cleaning and reshaping the Optum data using the R dplyr libraries. Using the R plm library on the Amazon cloud server, we examine the drivers of SGLT2 adoption. On the explanatory side of the model we have an array of lab values, patient demographics, insurance plan coverage and prior treatment trajectory.  Next we drill-down on insulin combination therapy. As an entire class, insulin is negatively associated with SGLT2 utilization, but the class coefficient masks a great of individual agent heterogeneity. Our presentation sketches a path to answering a number of significant research questions by accessing and modeling very large datasets using R.