Performance Impact of Immersion and Collaboration in Visual Data Analysis

Immersive Analytics is a recent field of study that focuses on utilizing emerging extended reality technologies to bring visual data analysis from the 2D screen to the real/virtual world. The effectiveness of Immersive Analytics, when compared to traditional systems, has been widely studied in this field's corpus, usually concluding that the immersive solution is superior. However, when it comes to comparing collaborative to single-user immersive analytics, the literature is lacking in user studies. As such, we developed a comprehensive experimental study with the objective of quantifying and analysing the impact that both immersion and collaboration have on the visual data analysis process. A two-variable (immersion: desktop/virtual reality; number of users: solo/pair) full factorial study was conceived with a mixed design (within-subject for immersion and between subject for number of users). Each of the 24 solo and 24 pairs of participants solved five visual data analysis tasks in both a head-mounted display-based virtual world and a desktop computer environment. The results show that, in terms of task time to completion, there were no significant differences between desktop and virtual reality, or between the solo and pair conditions. However, it was possible to conclude that collaboration is more beneficial the more complex the task is in both desktop and virtual reality, and that for less complex tasks, collaboration can be a hindrance. System Usability Scale scores were significantly better in the virtual reality condition than the desktop one, especially when working in pairs. As for user preference, the virtual reality system was significantly more favoured both as a visual data analysis platform and a collaborative data analysis platform over the desktop system. All supplemental materials are available at https://osf.io/k94u5/.