An essential activity in curriculum design is to specify the topics of the curriculum and the courses where those topics will be taught. Disciplines, such as Computing present several challenges in this regard, since the topics that students must learn tend to be fine-grained and highly interconnected. First, one must ensure that the most important topics of the curriculum are taught in at least one course. Second, for every topic taught, their prerequisites must have been covered previously in the same course or in a previous one. Third, courses must include topics that are highly cohesive and with minimal dependencies to topics taught in previous courses. To address the above challenges, this paper proposes a graph-based approach to analyze and design acurriculum, which also includes some Backward Design elements. Learning goals (desired results), topics, and courses are modeled as nodes in a graph. Prerequisite dependencies are modeled as edges. The relation between courses and topics are also modeled as edges. Graph analysis techniques are utilized to measure several aspects of a curriculum. Edges between topics are utilized to verify consistency between topics and prerequisite and corequisite relations between courses. Course-topic edges are used to calculate topic coverage of the curriculum. Topological sorting and course-topic relations are utilized to automatically generate the draft of course syllabi. We also describe the results of a real-life application and argue that this approach is essential to make visible and verify the overall structure of a curriculum.
