Big Data in the Science of Learning

The low cost of mobile computing devices, coupled with cheap and elastic cloud-based data storage and computing, along with advances in the fields of artificial intelligence and machine learning, has launched a data revolution in the science of learning. How do we gain knowledge about how people learn from data on hundreds of thousands of students and their teachers collected across extended time frames? This chapter considers this question, starting with a word about the data itself. How to make sense of all these data? The chapter briefly considers some of the key research areas and computational techniques used. The key research areas and computational techniques include: formative assessment; supervised, semisupervised, and unsupervised learning; relationship discovery; sequence modeling; content analysis and natural language processing; and multisensory, multimodal modeling. The chapter finally turns to four case studies from authors' lab to illustrate a broad range of research goals, data, and methods. (PsycInfo Database Record (c) 2024 APA, all rights reserved)