This work explores the implications of using thermal channels alongside physical tokens on a glove to create a wearable data physicalization that offers both public and private data displays. Encoding data using thermal sensations offers privacy and can complement a physical-visual display by representing one additional dimension. We provide thermal feedback using a nichrome wire, with temperature regulated by transistors controlled by a microcontroller. 3D-printed detachable structures (tokens) were developed to create a physical-visual display alongside the private thermal display. Our work also discusses different ways of representing human-centred data thermally, as well as real-world use cases such as self-reflection, empathy building and collaborative sensemaking.

While visual channels (e.g., color, shape, size) have been explored for visualizing data in data physicalizations, there is a lack of understanding regarding how to encode data into physical material properties (e.g., roughness, hardness). This understanding is critical for ensuring data is correctly communicated and for potentially extending the channels and bandwidth available for encoding that data. We present a method to encode ordinal data into roughness, validated through user studies. In the first study, we identified just noticeable differences in perceived roughness from this method. In the second study, we 3D-printed proof of concepts for five different multivariate physicalizations using the model. These physicalizations were qualitatively explored (N=10) to understand people’s comprehension and impressions of the roughness channel. Our findings suggest roughness may be used for certain types of data encoding, and the context of the data can impact how people interpret roughness mapping direction.

Parametric Design enables designers to formulate and explore new ideas through parameters, typically by manipulating numerical values. However, visualising and exploring the design space of an established parametric design solution is natively difficult through desktop displays for designers due to screen space constraints and requiring familiarity with visual-language programming interfaces. Thus, we sought to explore Virtual Reality (VR), inspired by Natural User Interfaces (NUI), to develop and explore new interfaces departing from traditional programming interfaces, that could complement the spatial and embodied affordances of contemporary VR devices. Informed by two industry-led focus groups with architects we developed and examined the usability of three different interfaces: 1) Paramaxes , an axes-based interface that allows designers to distribute and manipulate parameter visualisations around them in physical space; 2) ParamUtter , a Voice-based User Interface (VUI) that allows designers to manipulate parameter visualisations through natural languages; 3) Control Panel , which presents the parameters as sliders in a scrollable pane and acts as baseline comparison. We ran an exploratory study with experts and found that the Control Panel was ultimately the preferred interface for a design manipulation task. However, participants commented favorably towards qualities in the unconventional interfaces, with ParamUtter scoring highest in System Usability Scores (SUS), and participants valuing the potential of using physical space to explore design spaces with Paramaxes .

Existing situated visualization models assume the user is able to directly interact with the objects and spaces to which the data refers (known as physical referents). We review a growing body of work exploring scenarios where the user interacts with a proxy representation of the physical referent rather than immediately with the object itself. This introduces a complex mixture of immediate situatedness and proxies of situatedness that goes beyond the expressiveness of current models. We propose an extended model of situated visualization that encompasses Immediate Situated Visualization and ProxSituated (Proxy of Situated) Visualization. Our model describes a set of key entities involved in proxSituated scenarios and important relationships between them. From this model, we derive design dimensions and apply them to existing situated visualization work. The resulting design space allows us to describe and evaluate existing scenarios, as well as to creatively generate new conceptual scenarios.

Recent research in immersive visualisations has explored the use of physical 3D models together with virtual data visualisations, with a simple data encoding. However, little is known about how this technique could be extended towards more complex multivariate data with multiple charts. We present augmented scale models, immersive visualisations that place charts of multivariate data via Augmented Reality (AR) registered to physical 3D models. We identified two main factors for presenting AR charts in the limited display space around the models: 1) how charts are laid out (Slides vs Dashboard), and 2) how the chart views are arranged in the 3D space (On Scale Model, On Table, On Virtual Board). In a within-subject user study, we evaluated these two design considerations. We found that chart layout and view arrangement do not affect task error but do vary in response time. Dashboard and Slides perform equally well on simple tasks that require comparison of a single chart across scale models. However, in more complex tasks such as comparing multiple charts within a scale model or across several scale models, Dashboard shows no decrease in time performance, while Slides’s time performance decreases significantly. We also found that On Scale Model has the fastest performance, has good chart-scale model integration, and supports charts comparison well. On Table and On Virtual Board on the other hand, show a trade-off between the ability to support charts comparison across models and the chart-scale model integration.

Data physicalisations afford people the ability to directly interact with data using their hands, potentially achieving a more comprehensive understanding of a dataset. Due to their complex nature, the representation of graphs and networks could benefit from physicalisation, bringing the dataset from the digital world into the physical one. However, no empirical work exists investigating the effects physicalisations have upon comprehension as they relate to graph representations. In this work, we present initial design considerations for graph physicalisations, as well as an empirical study investigating differences in comprehension between virtual and physical representations. We found that participants perceived themselves as being more accurate via touch and sight (visual-haptic) than the graphical-only modality, and perceived a triangle count task as less difficult in visual-haptic than in the graphical-only modality. Additionally, we found that participants significantly preferred interacting with visual-haptic over other conditions, despite no significant effect on task time or error.

In augmented reality (AR), gesture/hand-based interactions are becoming more common place over tangible user interfaces (TUIs) and physical controllers. Major concerns regard portability and battery-life with TUIs and physical controllers in an AR environment. For a TUI or physical controller to be usable ”on the go” and in the field, they need to be compact such that they can be easily deployable. Subsequently, it is desirable to be low-energy or powerless, such that the device remains functional after long periods. In this paper, we present our initial design towards a powerless, passive controller that leverages the optical camera of a AR head-mounted display (such as the Hololens or Magic Leap) for tracking. We discuss related work, design motivations, high and low fidelity designs, opportunistic/adaptable input modalities, and use cases. We conclude with propositions for future work.

This paper presents development on a conceptual method to remap supernumerary limbs using Virtual Reality (VR) as a platform for experimentation. Our VR system allows users to control a third arm through their own limbs such as their head, arms, and feet with the ability to switch between them. To realize and experiment with our remapping method, we used the Oculus Rift in conjunction with OptiTrack to track users in a room-scaled virtual environment. We present some initial findings from a small pilot study and conclude with suggestions for future work.

Robot assistance can reduce the user's workload of a task. However, the robot needs to be programmed or trained on how to assist the user. Virtual Reality (VR) can be used to train and validate the actions of the robot in a safer and cheaper environment. In this paper, we examine how a robotic arm can be trained using Coloured Petri Nets (CPN) and Partial Least Squares Regression (PLSR). Based upon these algorithms, we discuss the concept of using the user's acceleration and rotation as a sufficient means to train a robotic arm for a procedural task in VR. We present a work-in-progress system for training robotic limbs using VR as a cost effective and safe medium for experimentation. Additionally, we propose PLSR data that could be considered for training data analysis.