Data analytics is a pervasive and powerful set of methodologies that allows researchers to gain insights into processes and behaviors. It does so by combining and processing a wide array of trace data that are made available by the internet and mobile technology. This is true also in the field of educational research, where the development of digital tools to support pedagogical activities offers new opportunities for studying learning. In turn, by collecting data about the usage of learning technology, we can better understand and reshape the contexts in which learning takes place.
Driven by the interest in data intensive research and stimulated by the challenges of applying data analytics to educational research, The Research Laboratory for Digital Learning has been working with the SpotOn project, an initiative of The Ohio State University, sponsored by the College of Education of Human Ecology. SpotOn’s mission is to “support PreK-12 teachers, administrators, and content decision-makers in the evaluation and selection of those digital resources that best meet their instructional requirements and the learning needs of their students.”
In this effort, the data analytics team in our research laboratory has been develping analytics and algorithms aimed at the implementation and deployment of an electronic recommendation engine that can provide customized recommendations of digital educational content to educators. As an online tool, automatic recommendations makes use of the traces we leave of the interactions with a website’s contents and helps us organize those contents along the most useful dimensions: Amazon’s recommendations help us filter through its vast catalogue to find our next purchase; Netflix suggestions present us the titles that are most likely to interest us. In our case, instead, we have developed and implemented a hybrid recommendation filter that will allow us to calculate the similarity among resources available on spotonreviews.org, using educators’ evaluations and the metadata available from publishers.
The first version of our engine is live and can be seen at the bottom of many resource pages, after the product’s description. At the moment, we are continuing to optimize its output. As the website’s user base grows and as trace data accumulate, the collaborative filtering algorithm at the core of our engine will allow us to improve our analytics, a result that we are happy to share back with the project team and all the visitors of the SpotOn website.
The development team members include Gennaro Di Tosto, Young Suk Cho, Lin Lu and Sheng-Bo Chen with the supervision of Prof. Kui Xie.