The amount of data extracted from learning experiences has grown at an astonishing pace both in depth due to the increasing variety of data sources, and in breath with courses now being offered to massive student cohorts. However, in this emerging scenario instructors are now facing the challenge of connecting the knowledge emerging from data analysis with the provision of meaningful support actions to students within the context of an instructional design.

The half-day session comprises of a set of hands-on activities with prepared material. Attendees are required to bring their own laptop. The nature of the topic will appeal to researchers and practitioners in the areas of learning analytics, educational data mining, instructional design, and any academic interested in improving the timeliness and relevance of feedback and support for their students. Attendees are not required to have experience with any programming language but need to be familiar with the conventional data management procedures required in a course with a large number of students such as for example combining scores and observations from multiple sources.

To guarantee a productive session, attendance will be limited to 30 people.


The objectives of the tutorial are to:

  • Understand the use of exploratory data analysis to summarize data sets derived from a learning experience.

  • Identify student support actions to be deployed while the experience is being delivered.

  • Define a set of low latency actions (those reaching students within a day) that are personalized to each student.

  • Express the connection between data and actions in a formalism suitable to be deployed at scale.


The tutorial complements other initiatives deployed in previous editions of LAK as well as other conferences focusing on educational technology such as ECTEL. The emphasis of this session is not on using specific data-mining methods, but on articulating the connection between the knowledge extracted from those tools and specific actions to support students.


The agenda for the session is as follows:

  1. The case for using data to provide personalized feedback to students (45 minutes).

  2. Scenario 1: Providing feedback on automatically graded assessments (45 minutes).

  3. Scenario 2: Encouraging student engagement in real time (60 minutes).

  4. Scenario 3: Using predictive models to suggest study strategies (45 minutes).


Tutorial Organizers

  • Abelardo Pardo, The University of Sydney
  • Roberto Martínez-Maldonado, University of Technology Sydney (UTS)
  • Simon Buckingham Shum, University of Technology Sydney (UTS)
  • Simon McIntyre, University of New South Wales, Autralia
  • Dragan Gasevic, The University of Edinburgh
  • George Siemens, The University of Texas at Arlington
  • Jurgen Schulte, University of Technology Sydney (UTS)