Brief Overview of the Paper In this paper, the authors discuss the adaptation of a commonly used visualization design pattern, known as small multiples displays, to immersive interaction spaces. Small multiples displays involve representing multiple data sets using the same visualization style in a tiled display, allowing for easy comparison.

Contributions Beyond Previous Work While previous studies have suggested different designs for 2D and 3D displays, this paper goes a step further by exploring the design space and evaluating small multiples in immersive environments.

What I Like About It I appreciate the wide range of applications that can be derived from the insights provided in this paper. In my own work for IEEE VR, I explored presenting miniatures in 3D space, and this paper’s exploration of different orientations for miniatures aligns with my research interests. I wish I had come across this paper earlier to learn about the design principles it presents. Additionally, the experiment conducted by the authors to test the placement of visualizations behind the user is valuable, as it helps determine if the limitations observed in the real world also hold true in virtual reality.

What I Don’t Like About It One aspect I find lacking is that the authors did not reorder the tasks in experiment 2. Task types may affect the learning speed of users, so reordering should have been considered here.

What I Think Should Have Been Done Differently I believe the authors should have cited Steve’s paper from 1990, titled “Worlds within Worlds.” This paper likely explored similar visualization techniques for presenting 4D data.

What I Think Should Be Done Next To further improve the study, I suggest reordering more variables for different participants. The learning effect observed in section 6.2, where the difficulty becomes harder to rely on, indicates the need for additional investigation. Additionally, the possibility of a learning effect in section 7.2 cannot be ruled out. Furthermore, the authors could explore additional design spaces, such as designs where each row is not straight but bent, or designs where visualizations are scaled differently (e.g., shrinking as they go further).

4D Data Analysis Applications Listed

  • 4D UI
  • The authors envision applications for immersive small multiples in domains that rely on the exploration of 3D data. For example, analyzing changes in aircraft trajectories above an airport over different time periods can reveal patterns about the efficiency of airspace utilization and the risks of collision. Another relevant domain is Building Information Modeling (BIM), which involves managing a facility’s digital information assets. Building managers can benefit from comparing temperature sensor readings and energy consumption over time to identify trends. Analysts may also be interested in data that does not have a physical spatial embedding or abstract quantitative data. For instance, using 3D bar charts to compare wealth and productivity statistics across different populations.