Accounting for complex interactions between contextual variables and learners' individual differences in aptitudes and background requires building the means to connect and access learner data at large scales, across time, and in multiple contexts. This paper describes the ASU Learning@Scale (L@S) project to develop a digital learning network platform with the capacity to connect, access, and examine undergraduate student data and courses. Arizona State University (ASU) collectively serves over 100,000 undergraduate students and 40,000 K-12 students within traditional in-person courses, online and blended courses, Earned Admission, and ASU Prep Online (K-12). Combined, these programs at ASU provide access to the large-scale data needed to improve the generalizability of learning sciences research. Foremost, we are challenged by this wealth of data at ASU currently being housed in multiple locations, with varying restrictions, and means of access. The ASU L@S team is developing the data and policy infrastructure necessary to enable data access while supporting multiple types of research (e.g., historical data analyses, rapid A/B testing, efficacy studies, design studies). Our objective is to render learner data and educational contexts available to both internal and external researchers, and to facilitate researchers' ability to conduct research more efficiently on ASU student learning while maintaining the safety of all stakeholders' personally identifiable information. Here we describe the three primary datasets currently being compiled within the ASU L@S data warehouse (i.e., Student Profile, Student Trajectory, Course Profile) and two datasets leveraging natural language produced by learners in various contexts: discussion board posts and written assignments. Combining and integrating these datasets within a single data warehouse sets the stage to enable impactful embedded research at ASU that enhances student outcomes and in turn contributes to theories of how people learn.