Gallery — studies built with EasyStudy

Real studies deployed with EasyStudy, as evidence that it scales from a class exercise to published research. See also Studies built with EasyStudy for the full, regularly updated list of peer-reviewed papers.

Featured studies

Rows or Columns? Minimizing presentation bias when comparing multiple RS

• What: measured positional/presentation bias across page layouts when comparing several recommenders.
• Venue: SIGIR '23 · DOI:10.1145/3539618.3592056
• Plugin/branch: layoutshuffling (alternates layouts during the study).

Controllable multi-objective recommenders (slider UI)

• What: compared single- vs multi-objective RS; a slider UI let users set propensity toward relevance/novelty/diversity. Yielded the SM-RS 2.0 user-perceived-qualities dataset.
• Venue: ACM TORS · DOI:10.1145/3754459 · dataset osf.io/wsakx
• Plugin/branch: feature/multiObjectivePlugin / slidershuffling.

EasyStudy (the framework demo)

• What: the original within-user comparison of EASE / MF / kNN baselines on MovieLens & goodbooks.
• Venue: RecSys '23 · DOI:10.1145/3604915.3610640
• Plugin/branch: fastcompare (feature/recsys2023).

Data-visualization user studies (domain transfer)

• What: EasyStudy adapted outside recommender systems, for data-visualization experiments — concrete proof the framework is domain-generic.
• Venue: IEEE — ieeexplore 10854212 · and Springer chapter 10.1007/978-3-031-49368-3_14

Teaching at Charles University (NDBI021 / NSWI166)

• What: used in RS courses both to demo algorithms live and to evaluate students' solutions.
• Branch: ndbi021 (lightweight, minimal dependencies).