Researchers on sentencing often lament the lack of available information in their analysis.  Therefore, several prominent literature reviews at the beginning of the decade (Baumer, 2013; Ulmer, 2012) set the discovery of data as one of the top priorities for researchers.  While there are many practical constraints associated with obtaining new data, another potential way to move forward is to further explore the potential of existing datasets, particularly those large administrative datasets.  This talk, consisting of three early-stage projects, showcases some promising and less promising directions of utilizing a large dataset for research on sentencing and criminal careers.  The studies seek to answer the same question: how far can we go with the current generation of data?