Peer recommender for MOOCs (GroupMOOC)

MOOCs
recommender systems
pattern: static artifact
Pattern (b) — embed pre-built artifacts; the site executes no code at all.
Published

1 May 2017

Modified

20 June 2026

ImportantIntegration pattern (b): static artifact

This page contains no executable code. The figures and tables are produced in the project’s own repository (by its own scripts or CI) and then copied into this site as static files under assets/img/. Building the website therefore touches no Python environment whatsoever — the most decoupled option, ideal for heavy or dormant projects.

Refresh the artifacts when the project changes:

scripts/sync-artifacts.sh /path/to/groupmooc/figures assets/img
git add assets/img && git commit -m "build(projects): sync GroupMOOC figures"

What it is

GroupMOOC studied how connecting learners to one another can reduce attrition in MOOCs. Within it, we built a peer recommender that suggests high-potential contacts in real time, evaluated in a controlled study on the Gestion de Projet MOOC (N ≈ 8,700), with Hugues Labarthe, Rémi Bachelet and Kalina Yacef.

Result (static figure)

Figure 1: Completion and engagement were higher for learners who received recommendations than for the control group.

The figure above is a plain image: it was exported by the project and committed to this repository. Nothing is computed when the site is built.