# From Learning Traces to Teaching Insights: Dodona, Dolos, and What Student Data Reveals ![Learning data and analytics pink and orange gradient logo.](/sites/g/files/omnuum12666/files/styles/hwp_5_4__480x385/public/2026-10/250624_LearningDataAndAnalytics-Logo002.png?h=e27e3a27&itok=Fc2set0j) #### calendar\_today Date and Time **October 27, 2026** 02:00PM - 02:45PM EDT #### pin\_drop Location **Zoom** Please Register for this event! [ Register Today! arrow\_circle\_right ](https://harvard.zoom.us/meeting/register/J6Gbq4FHQ1KxuBjPF_Yh-w) Hosted by the Learning Data and Analytics (LD&A) HILT Affinity Group Join us for a session with Peter Dawyndt (Professor of Computer Science, Ghent University), Bart Mesuere (Professor of Data Visualization, Ghent University), and researchers Rien Maertens and Michiel Lachaert, the team behind Dodona and Dolos. Dodona, developed since 2016 and currently a university spin-off, supports programming, statistics, and data science courses and serves more than 100,000 learners. The platform integrates with Dolos, an open-source plagiarism detection tool for source code designed for teachers, with visualizations that can be shown live in class using pseudonymized data. The team's philosophy is simple: "Don't collect data just to collect data." When learners are active, they leave traces of how they learn, and those small traces hold real insight for teachers. The question is what you do with it. The team will share what the data shows, including: - Early prediction of student success. By week 3, submission data predicts pass or fail outcomes with high accuracy, early enough to intervene. The finding was replicated at a university in Finland. - AI and learning outcomes. In last year's course data, about 70% of students who never used AI while practicing passed the final exam, compared with 28% of those who used AI at least once. - Classroom interventions that work. How one well-timed intervention using live plagiarism visualizations prompts genuine student reflection, and why restraint matters. - Where plagiarism detection fits in the wider story of teaching, from active learning through homework to exam preparation. Together, the two tools show the full arc: Dolos identifies where learning is breaking down, and Dodona is where teachers act on that insight, in the classroom and in the design of the course. Whether you teach, design courses, or work with learning data, you will leave with concrete research findings and practical approaches you can connect to your own context. 45 minutes: introduction, presentation, demo, and 10 to 15 minutes of Q&A. **Questions?** Please reach out to the LD&A Co-chairs: **Allain Williams** [allain\_williams@harvard.edu](mailto:allain_williams@harvard.edu); **Alyshia Keys-Harris** [alyshia\_keys-harris@gse.harvard.edu](mailto:Keys-Harris,alyshia_keys-harris@gse.harvard.edu); **Kama Lord** . Save: [ Add to calendar calendar\_today ](https://www.vpal.harvard.edu/node/1931816/event-feed.ics) Copy link link