Due to various historical events, the Aviation Engineering Directorate (AED) of the United States Army has a unique, large data set describing the mechanical health of rotorcraft systems. This data set includes detailed information regarding non-critical failures and wear over a significant period and number of aircraft, each of which is instrumented to take measurements of mechanical vibrations and other parameters every flight. Attempts to utilize this data led AED to investigate the efficacy of machine learning and knowledge discovery from data (KDD) techniques. This paper outlines a tool–termed the Crawler–which AED developed to automate the process of creating diagnostic classifiers, and its application to two specific problems of interest: improving the usability and performance of a deployed gearbox health classifier, and rapidly developing a model to search sensor data for a newly identified fault mode.