In the past decades, the amount of information available to law enforcement agencies has increased significantly. Most of this information is in textual form, however analyses have mainly focused on the structured data. In this paper, we give an overview of the concept discovery projects at the Amsterdam-Amstell and police where Formal Concept Analysis (FCA) is being used as text mining instrument. FCA is combined with statistical techniques such as Hidden Markov Models (HMM) and Emergent Self Organizing Maps (ESOM). The combination of this concept discovery and refinement technique with statistical techniques for analyzing high-dimensional data not only resulted in new insights but often in actual improvements of the investigation procedures.