• Research

    • Learner Modeling & Learning Analytics at Scale

Learner Modeling & Learning Analytics at Scale


The Online Education Research program at TU Delft targets the elements of online education that align with Delft research strengths and allow us to support the next generation of online education – not just in Delft (as EdX partner we are at the forefront of the current drive towards MOOCs), but beyond.  With bringing education online and doing so at unprecedented scales,, the role of technology in online education and its research is showing a paradigm shift. The nature and volume of learning data, the interaction between students and educators, and the interaction between all stakeholders in education change, and as a consequence the relevant research becomes more data-driven and computational.

Our Mission

The Web Information Systems group contributes to this research team and focuses on Learner Modeling and Learning Analytics.

Analyzing the large amount of digital traces learners leave within and outside the online learning environment such as EdX offers the possibility to adapt the environment itself, the teaching material and the manner of conveying knowledge to the individual learners’ abilities and preferences. Relying on research methodologies developed across diverse fields such as educational psychology, human-centered design, data science and big data processing, we investigate three main questions within the learning analytics theme:

  • How is the design of the learning environment related to students’ engagement and achievement in open online higher education?
  • Digital learning environments do not exist in isolation, they are embedded in the rich infrastructure of the Web. Learners regularly seek out additional materials available elsewhere on various portals, educational websites, etc. Moreover, we can seek out the learners’ themselves on the social Web, inferring and building knowledge about them.  To what extent do these external sources enable us to explain and improve students’ engagement and achievement?
  • Commonly, to analyze educational traces, adhoc pipelines are used that are specific to a particular use case. We envision to create a generic platform for conducting analytical tasks with large-scale traces that allow us to build not only research prototypes but also actual applications that make use of the analytics results.


As Web Information Systems group we are not only conducting data-driven research, we also design and develop software. This project will lead to two concrete outcomes in terms of software:

  • Delft-X-Analytics, our Web-based MOOC dashboard that allows stakeholders at TU Delft (lecturers, instructional designers, etc.) to investigate the impact our TU Delft MOOCs have on students across a wide range of metrics.
  • U-Sem-L, the specialisation and extension of U-Sem, our modelling framework (developed during an earlier project) for applications in Massive Open Online Learning.

Researchers Involved

Several researchers of the WIS group contribute to this research line:

  • Guanliang Chen, a PhD student, who is working on learner analytics and the implementation of components for U-Sem-L.
  • Claudia Hauff, an Assistant Professor, whose main focus is the embedding of online learning into the wider Web.
  • Geert-Jan Houben, head of the WIS group, whose expertise in user modeling and Web systems will guide this project.

Prior Works

[Click to Enlarge]

Large-scale data analysis of educational traces is recent phenomenon, which has gained
sufficient attention to warrant a dedicated conference, ACM Learning At Scale.

Most prior works in this area (graphically depicted on the left) have
investigated various relationships between learners and the learning
environment, the instructors and the courses themselves.

It is our strong belief that the inclusion of the wider Web in the picture and
the push for personalized learning at scale will bring about new insights and
a better learning experience.

Recent Activities

  • [Upcoming]: Daniel Davis will join WIS in the summer as PhD student, investigating the learners’ engagement with online learning environments
  • [March 2015] We co-organize the Delft Data Science Seminar on Speeding Up the Online Learning Curve. All talks can be found online.
  • [March 2015] Guanling Chen starts as PhD student on Learning Analytics @ WIS.
  • [January 2015] Claudia Hauff presents a pitch on Social data for lifelong learning at TU Delft’s 173rd Dies Natalis. Have a look at the slides.
Be Sociable, Share!
  • Twitter
  • Facebook
  • LinkedIn
    © 2011 TU Delft