As students use learning technology they leave behind traces of their interactions that are beneficial in understanding the teaching and learning environment. Using a combination of machine learning and information visualization techniques, these traces can be aggregated, summarized, and represented such that instructional experts can gain insight into the activities of learners. These insights can then be used to compare cohorts of learners with one another, change teaching methods and techniques, or modify the availability of tools in the learning environment. This seminar will discuss the results of two different investigations into how learner traces can be used to influence instructional experts. The first of these investigations used visualization methods in the iHelp Discussions system. The iHelp Discussions system was used by more than 6,000 students in the Department of Computer Science between 2004 and 2010. It is an asynchronous discussion forum for peer help, and novel sociogram-based visualizations were employed to demonstrate to instructors the interactions of learners (both with the system and with one another). These visualizations were effective in helping instructors identify weaknesses in their pedagogical approaches, which led to a change in their teaching practice. The second investigation applied unsupervised machine learning to learner traces created by the Recollect lecture capture system. The Recollect system was used by nearly 4,300 learners across multiple disciplines at the University of Saskatchewan from 2008-2011, and served as a basis for the open source Matterhorn system which has since been deployed campus-wide. Relationships between academic achievement and patterns of usage of Recollect were discovered across a number of STEM courses, and this talk will summarize these results as well as describe the opportunity that exists for extending learning analytics to other common learning technologies.