Data, Mark of a New Era

Abstract

Data is at the heart of learning analytics; collecting and utilizing the appropriate data is key to having useful and actionable outcomes from learning analytics models. In this chapter, we focus on using quality data as an essential aspect of developing models that use learning analytics. We first introduce data types and then discuss the sources of data that are useful for use in learning analytics models. Finally, we provide detailed examples of these different data types in use in learning analytics models. These examples fall broadly under the categories of prediction, clustering, and rule mining. Since most learning analytics models make use of multiple types of data in developing a robust and accurate model, we do not separate out examples based on data type alone. Instead, our examples bring together both data types and the intended model use. For example, in our discussion of prediction using learning analytics, we discuss learning analytics for improving student retention as well as using novel forms of data, like affect data, in predictive learning analytics models. Learning analytics is used not only in traditional reporting, evaluation, and decision-making processes but also for the new paradigm of data driven learning such as adaptive education and intelligent tutoring. We also provide real world case studies, like Squirrel AI Learning and ASSISTments. Quality data forms the foundation for any good learning analytics model; this chapter seeks to underscore considerations the importance of data in model development.