Scaling Up and Zooming In: Big Data and Personalization in Language Learning.

From its earliest days, practitioners of computer-assisted language learning (CALL) have collected data from computer-mediated learning environments. Indeed, that has been a central aspect of the field from the beginning. Usage logs provided valuable insights into how systems were used and how effective they were for language learning. That information could be analyzed to improve instructional design and delivery. Maintaining learning histories and personal profiles of individual learners enabled a program to adapt the delivery of learning materials to the record of student performance. Given a limited number of users working within a single system, the data generated could be collected and analyzed easily, using simple methods and tools such as spreadsheets and basic data models. The situation today is quite different from that scenario. Learners are likely to be using multiple online tools and services, all of which may be recording data. That includes general use software and services such as Facebook and Google, as well as mobile devices. If they are university students, they are likely to be generating data points through a learning management system (LMS) as well as from other university-level systems. The vast amount of information collected today from our use of online tools and services provides a huge storehouse of information that can be mined to provide both general usage trends and individualized reports. This big data offers valuable teaching and learning insights. In this column, we will be looking at what this may mean in language learning. That will include discussion of the emerging field of learning analytics, the use of learner models, and the opportunities afforded by data tracking for personalized learning.