This chapter explores the challenges of learning analytics within a virtual learning infrastructure to support self-directed learning and change for individuals, teams, organizational leaders, and researchers. Drawing on sixteen years of research into dispositional learning analytics, it addresses technical, political, pedagogical, and theoretical issues involved in designing learning for complex systems whose purpose is to improve or transform their outcomes. Using the concepts of 1) layers - people at different levels in the system, 2) loops - rapid feedback for self-directed change at each level, and 3) processes - key elements of learning as a journey of change that need to be attended to, it explores these issues from practical experience, and presents working examples. Habermasian forms of rationality are used to identify challenges at the human/data interface showing how the same data point can be apprehended through emancipatory, hermeneutical, and strategic ways of knowing. Applications of these ideas to education and industry are presented, linking learning journeys with customer journeys as a way of organizing a range of learning analytics that can be collated within a technical learning infrastructure designed to support organizational or community transformation.