Topic outline

  • General

    Course description

    Contents: Educational Data Mining (EDM) as a field focuses on developing methodologies and tools to explore and analyze data gathered from educational settings. The roots of EDM are in constructing, studying, and improving systems that guide and teach students, with the (implicit) goal of improving offered education. Used methods and practices include various forms of data mining, including machine learning. Depending on your interests, you can focus on, e.g., supervised methods such as predicting student success, unsupervised methods such as clustering student data to identify student subpopulations and behaviors, association rule learning to find if-then patterns in data. You can also suggest a topic of your own within the scope of the seminar.

    Assessment: Project work and presentation, Pass/Fail


    Learning objectives for the seminar:

    • Understands what educational data mining is.

    • Understands core challenges in the educational data mining domain.

    • Improves scientific and technical writing and presentation skills.

    • Gains hands-on experience from constructing, reporting, and evaluating an experiment using EDM methods.


    Implementation: During the first two weeks, the seminar jointly conducts a brief literature review of the educational data mining domain, focusing on recently published work. This is followed by planning, implementing, evaluating, and presenting a small-scale EDM experiment in groups. Datasets for projects will be provided by the instructors.


    Timeline: Meeting twice per week on weeks 1 and 2, once a week on weeks 3 - 7

    • Weeks 1 and 2, literature review

    • Weeks 2 and 3, planning project in groups

    • Weeks 3 to 6, project implementation and evaluation

    • Week 7, project presentations


    Possible topics and keywords to search for:

    • Extracting knowledge components

    • Knowledge tracing / Modeling students' knowledge

    • Plagiarism detection

    • Automatic hint generation

    • Predicting course outcomes / Predicting student dropouts

    • Characterizing students

    • Analyzing open-ended text responses

    • .. etc

    For the first meeting, please skim through the article Educational data mining and learning analytics: An updated survey (https://bookdown.org/chen/la-manual/files/Romero%20and%20Ventura%20-%202020.pdf). We'll discuss the article during the meeting in addition to going over practical details related to the course.

    Contact information:
    • Arto Hellas: arto.hellas@aalto.fi
    • Juho Leinonen: juho.2.leinonen@aalto.fi
    • Sami Sarsa: sami.sarsa@aalto.fi