The Uedu Educational Data Analytics Challenge (Uedu EDAC) is a recurring series of educational data science competitions. Think of Kaggle, but focused on educational data, learning analytics and precision education. There is one challenge cycle per semester and one annual theme per academic year, built around self-directed learning and automated platform assessment. You will complete the full pipeline of learning analytics → learner modeling → learning prescription: not just predicting, but proposing explainable, actionable learning interventions.

Three things set it apart: from prediction to prescription (we do not stop at a model score), transparent AI collaboration (generative AI is allowed, but must be honestly disclosed), and open-science governance (every challenge uses public datasets under their original licenses).

• Progression tiers: Explorer → Contributor → Expert → Master, accumulated across cycles and valid permanently.
• Digital badges and e-certificates: a Data Literacy badge and a completion certificate (both verifiable online and suitable for your CV).
• Hall of Fame: winners and the best work are permanently archived and showcased.
• Outstanding Fellows: at most 3 per cycle, given priority on the candidate list for follow-up research training.

Q: Will someone mentor me one-on-one?
No one-on-one mentoring is provided. Learning support consists of the starter notebook, the Uedu AI teaching assistant in the challenge course, and the course discussion board. This is a self-directed challenge.

Q: Can I form a team?
No, this is an individual challenge. Every participant works with the Uedu AI teaching assistant in the challenge course so that the judges can see each person’s own AI collaboration process.

Q: What is the challenge course?
After enrolling you are automatically added to the current challenge course (it cannot be searched or claimed; enrollment is the only way in). The course provides the Uedu AI teaching assistant (to help you write code, debug, interpret model results and develop learning prescriptions), challenge announcements, the dataset and starter notebook, and a discussion board.

Q: Can I use AI tools? What do the judges look at?
Yes, and we encourage you to use them well. Please do your AI collaboration for modeling and report writing with the Uedu AI teaching assistant in the challenge course. The organizers and judges can see these conversation logs; they are the basis for the “AI collaboration process” component (30%) of the overall score. Judges look at the quality of collaboration: whether you ask follow-up questions, verify the AI’s answers and make your own corrections, not how many conversations you had. Participants with no modeling-related AI conversations in the challenge course will lose a substantial share of the “AI collaboration process” score, and the judges will question them about their modeling and creative process at the final; that questioning time counts against the participant’s presentation time. If you also use external tools such as ChatGPT or Claude, you must disclose this in your prescription report; undisclosed use that is discovered leads to disqualification from the final.

Q: How do you make sure leaderboard scores are genuine?
With your final submission you must include the complete code (a notebook or .py file, compressed as a zip) that reproduces your final submitted result. The organizers re-run and verify the code of finalists; anyone whose submitted score cannot be reproduced is not ranked. The challenge dataset is derived from the public OULAD; linking it back to the original data to look up labels results in disqualification.

Q: Will my data be used for research?
Only with your consent. The enrollment page invites you to read the “Consent for Research Use of Platform Learning Process Data”. You can enroll whether or not you consent; it does not affect your eligibility, judging or ranking, and you can withdraw at any time. Judges viewing AI conversation logs is a challenge rule and is separate from research consent.

Q: What data does the challenge use?
Always international public educational datasets. This cycle uses OULAD: anonymized real learning records from The Open University (UK), released under CC BY 4.0, containing no personally identifiable information and no data from Uedu platform users. All derived files provided by the platform carry the original attribution.

Q: Does taking part affect my course grades?
No. This is an extracurricular, self-directed challenge with no connection to any course grade.

Q: How are Fellows selected?
Faculty select them according to pre-announced criteria (overall score, performance in the final and research potential), with at most 3 per cycle and the option to leave the title vacant. Fellow eligibility requires verification of an academic institution email address (e.g. .edu.tw in Taiwan, .edu in the US, .ac.jp in Japan, .ac.kr in Korea).