A data mining and visualization tool for the discovery of student trails in web-based educational systems is presented and described. The tool uses graphs to visualize results, allowing non-expert users, such as course instructors, to interpret its output. Several experiments have been conducted, using real data collected from a web-based intelligent tutoring system. The results of the data mining algorithm can be adjusted by tuning its parameters or filtering to concentrate on specific trails, or to focus only on the most significant paths.