This paper explores how student performance across the full set of multiple modular assessments of individual standards, which we refer to as mini-assessments, from a large scale, operational program of interim assessment can be summarized using Bayesian networks. We follow a completely data-driven approach in which no constraints are imposed to best reflect the empirical relationships between these assessments, and a learning trajectory approach in which constraints are imposed to mirror the stages of a mathematic learning trajectory to provide insight into student learning. Under both approaches, we aim to draw a holistic picture of performance across all of the mini-assessments that provides additional information for students, educators, and administrators. In particular, the graphical structure of the network and the conditional probabilities of mastery provide information above and beyond an overall score on a single mini-assessment. Uses and implications of our work are discussed.