ConstrucTED: Constructing Tailored Educational Datasets from Online Courses

Aymen Bazouzi, Zoltan Miklos, Mickaël Foursov, Hoël Le Capitaine

2024

Abstract

Researchers are actively involved in developing various systems to support education, including recommender systems. However, to create and evaluate such systems, they require rich and versatile datasets about educational content. At times, the available data proves insufficient, leading researchers to invest significant time in crafting personalized web scrapers for additional data retrieval. The generated datasets are often task-specific and may be time-consuming to adapt to future tasks. Additionally, researchers may encounter licensing issues when using courses from different providers. Furthermore, researchers prefer evaluating their methods through diverse tests, involving datasets with varying characteristics. However, this diversity is not commonly found in most available datasets, at least not explicitly so. To address these challenges, we introduce ConstrucTED, a tool built on top of Google APIs, enabling the efficient creation of custom educational datasets from YouTube playlists. This allows datasets to be tailored to specific characteristics such as a predetermined number of courses, coverage of specific topics, or courses from a particular university. ConstrucTED creates datasets from video course transcripts, providing a ready-to-use solution that significantly shortens the time required to create such datasets. The resulting datasets are versatile and suitable for tasks like classification and learning path creation.

Paper Citation

in Harvard Style

Bazouzi A., Miklos Z., Foursov M. and Le Capitaine H. (2024). ConstrucTED: Constructing Tailored Educational Datasets from Online Courses. In Proceedings of the 16th International Conference on Computer Supported Education - Volume 1: EKM; ISBN 978-989-758-697-2, SciTePress, pages 645-652. DOI: 10.5220/0012745000003693

in Bibtex Style

@conference{ekm24,
author={Aymen Bazouzi and Zoltan Miklos and Mickaël Foursov and Hoël Le Capitaine},
title={ConstrucTED: Constructing Tailored Educational Datasets from Online Courses},
booktitle={Proceedings of the 16th International Conference on Computer Supported Education - Volume 1: EKM},
year={2024},
pages={645-652},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012745000003693},
isbn={978-989-758-697-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 16th International Conference on Computer Supported Education - Volume 1: EKM
TI - ConstrucTED: Constructing Tailored Educational Datasets from Online Courses
SN - 978-989-758-697-2
AU - Bazouzi A.
AU - Miklos Z.
AU - Foursov M.
AU - Le Capitaine H.
PY - 2024
SP - 645
EP - 652
DO - 10.5220/0012745000003693
PB - SciTePress