The authors describe an innovative technique for using Classification and Regression Tree (CART) with student characteristics, including geographical data for analyzing student enrollment patterns between students who enroll within a community college district (stayers) and those who enroll outside the district (movers) at neighboring community colleges. This article describes a method for conducting a student flow analysis and demonstrates the method with an actual analysis at Gavilan College in Gilroy, California. This is a useful tool for research and planning professionals involved in enrollment management and marketing for community colleges at the program or discipline level. (Contains 1 figure and 4 tables.)