In line with developments in many areas of business and governance, where bureaucracies of all sorts are increasingly datafied for budgetary reasons and the additional possibilities for automated analysis, the Dutch Police started with so-called Intelligence-Led Policing. This development led to the creation of the Crime Anticipation System (CAS). This data-driven system tries to predict crimes with statistics based on three data sources: BVI (Central Crime Database), GBA (Municipal Administration) and CBS (Demographics from Statistics Netherlands). By analyzing the used data categories with a critical data studies approach, we will show that the epistemological question concerning predictive policing systems turns into an ontological one: how are living environments and police work mutually shaped and determined by data? We will argue that intelligence-driven policing is not only a qualitative shift, but also has its continuities, since already existing ideas and biases concerning suspects and crimes are reproduced in the information and system of CAS.