This website uses cookies and similar technologies to understand visitors' experiences. By continuing to use this website, you accept our use of cookies and similar technologies,Terms of Use, and Privacy Policy.

Jul 05 2017 - 07:50am
An overview on the state of Educational Data Mining

Opening Ceremony

Last week from June 24th to June 27th I attended the annual International Educational Data Mining conference in Wuhan China. Wuhan, China a less well-known city than Beijing or Shanghai, has a population of over ten-million. The city experienced a surge in development in 2006 and is continuing to add new subway lines and developments which we witnessed right outside of the conference location. Wuhan boasts the largest college and university student population (~1 million) in the country with the worlds largest education system. The sponsoring institution, the Central China Normal University (CCNU), also features the only national research lab for educational big data. This context, which echoed many of the comments made by the EdLab seminar presenters from Blue Elephant, was presented to us as part of the opening remarks of the conference.

Wuhan, Optics Valley area

I had the privilege of being able to present our very own EdLab  research and application of Gaussian Mixture Models on the Quuppa Data. Our research was accepted back in March and gathered much interest at the conference. While networking and talking with many of the attendees, researchers, industry leaders, and presenters, I learned our that our work on real-time location data was unique in the field in that it focused on "IRL" physical learning spaces as opposed to online learning platforms.  Since March we have completed and advanced further research in this field, by developing faster, more flexible, and more accurate analysis tools. Seeing the positive reaction to our initial research from other educators, researchers, and academics was encouraging for the future of our work.

Central China Normal University

The common goals for many of the research topics presented included: using data to better and more efficiently achieve learning outcomes, predicting success or failure, automation of online learning platforms, and developing learning analytics tools and models for educators. A model of learner behavior or user interactions, can serve as an easy to understand overview of a large set of data. In our own research, our method of clustering data, fit into the current popular methods and approaches used to conduct educational data mining.

Conference goers very interested in Topic Modeling and SNA analysis of Pressible - a research topic presented by EdLab's own Xiaoting Kuang

There were many challenges during the conference including limited access to internet. Despite our best efforts, the lack of access to services we take for granted such as Google proved challenging and it seemed the conference overwhelmed the hotel's internet capacity. Nevertheless, I came away from the conference with an expanded set of motivations, visions, and ideas which I hope to share with the EdLab team in the coming weeks.

Model of User Behavior on Online Course Lesson Video

I attended over 35 talks during the conference and attended an additional day of topic-focused workshops. In the following series of blog posts I will seek to provide an overview of some of some of the most popular topics in EDM as well as the implications for the work we are doing here at EdLab.

Find the conference proceedings here.

Educational Data Mining Topics: Assessment Automation

Educational Data Mining Topics: Collaborative Learning, Pedagogical Policies, and Educational Resources

Educational Data Mining Topics: Educational Resources and Affect and Engagement

Educational Data Mining Topics: Learner Behaviors

Educational Data Mining Topics: Poster Presentations

Educational Data Mining Topics: Ethics and Educational Data Mining

Our flight path

Posted in: ResearchTechnology|By: Alvaro Ortiz-Vazquez|577 Reads