@inproceedings{vinci2025,
author = {Al-Hazwani, Ibrahim and Mylaeus, Matthias and Mormocea, Daniela and Bernard, J\"{u}rgen},
title = {HUMMUS: Blending Data Humanism with Sequential Music Recommender Systems to Foster Explainability and Scrutability},
year = {2025},
isbn = {9798400718458},
publisher = {ACM},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3769534.3769578},
pdf = {https://dl.acm.org/doi/epdf/10.1145/3769534.3769578},
doi = {10.1145/3769534.3769578},
abstract = {Current music recommendation systems often lack transparency, preventing users from understanding the recommendations or effectively steering the algorithm. We present HUMMUS, an interactive collaborative music sequential recommender system that applies Giorgia Lupi’s Data Humanism principles to combine algorithmic transparency with human-centered design. HUMMUS visualizes songs as flowers, where petals represent audio features, and connecting lines reveal recommendation relationships. Real-time voting mechanisms during natural pauses in social interaction enable collaborative decision-making between humans and recommendation algorithms. Our mixed-methods evaluation, involving 19 participants, demonstrates that humanistic design principles enhance transparency, user engagement, and collaborative decision-making while maintaining the quality of recommendations. This work contributes to the intersection of critical visualization and explainable AI by demonstrating how Data Humanism can guide human-centered recommendation systems.},
booktitle = {Visual Information Communication and Interaction},
articleno = {53},
numpages = {5},
keywords = {Data Humanism, Explainable AI, Sequential Music Recommendations, Human-Centered AI, Human-AI Collaboration},
teaserpage = {1},
topics = {Explainable and Trustworthy AI, Design Studies and Applications},
code = {P031},