Rush: repeated recommendations on mobile devices

We present rush as a recommendation-based interaction and visualization technique for repeated item selection from large data sets on mobile touch screen devices. Proposals and choices are intertwined in a continuous finger gesture navigating a two-dimensional canvas of recommended items. This provides users with more flexibility for the resulting selections. Our design is based on a formative user study regarding orientation and occlusion aspects. Subsequently, we implemented a version of rush for music playlist creation. In an experimental evaluation we compared different types of recommendations based on similarity, namely the top 5 most similar items, five random selections from the list of similar items and a hybrid version of the two. Participants had to create playlists using each condition. Our results show that top 5 was too restricting, while random and hybrid suggestions had comparable results. Author Keywords Interaction technique, mobile; recommender systems ACM Classific...

• Type: Conference
• Year: 2010
• Where: IUI
• Authors: Dominikus Baur, Sebastian Boring, Andreas Butz
• Keywords: Continuous Finger Gesture, IUI 2010, Keywords Interaction Technique, Similar Items, Software Engineering