The Cross Industry Standard Process for Data Mining (CRISP-DM) is the leading published methodology for Data Mining (DM), and by extension, Predictive Analytics (PA). I use it routinely as I lead PA projects and when I teach appiled DM and PA courses. It was the subject of three KD-Nuggets polls, in 2002, 2004, and 2007) and nearly half of the responders stated they used it as the main methodology for DM (http://www.kdnuggets.com/polls/2007/data_mining_methodology.htm). There are six stages in the CRISP-DM process, including Business Understanding, Data Understanding, Data Preparation, Modeling, Evaluation and Deployment. There are many excellent books describing five of these