Learning any time, anywhere: Big educational data from smart devices

Abstract

For many people, especially young people, a smartphone is a constant companion. Mobile apps which allow individuals to use a smart device to enhance their learning have the potential to be very useful for mastering basic educational material. In order to evaluate and enhance the effectiveness of such applications when deployed at large scale, an infrastructure designed specifically for the collection of educational analytics data from such mobile apps is required. We detail here a set of applications and their associated infrastructure which was developed to allow students in courses using digital textbooks to enhance their knowledge of the basic course content anywhere and anytime by using their smart device to do spaced practice of the knowledge components of a course. The power of current smart devices allows the entire application, including content and adaptive algorithm to be hosted and run locally on the user’s smart device, so it functions fully even when no network connection is available. The infrastructure for the collection and analysis of the educational analytics data is entirely cloud-based, using AWS S3 for data collection and storage, and the Apache Spark parallel computing framework for data analysis. Thus, the entire system requires only laptop computers for the mobile developers who create the applications and this is also sufficient for the learning scientists who analyze the data. Both the data collection system and the data analysis system can scale to handle the data from many millions of users with no modification to their architecture. Similar architectures are now used for the Internet of Things (IOT) but have not yet been widely used for educational applications. These applications have currently been deployed to thousands of users’ smart devices and analytics data is being received from these users’ smart devices from a wide range of locations on several continents. In our highly connected world, this type of application will become much more common. We describe here the type of infrastructure, security, and analytic methods needed to use these apps to advance learning and learning science.