Empowering Users with Composite Visualizations

As data grows more complex and abundant, the need for better ways to visualize it has increased. To get a clear understanding of data, it's often helpful to combine multiple visual representations into one. This is what we call composite visualization. However, creating these visualizations has usually required skilled experts who work behind the scenes, which can make the process feel disconnected for the average user.

This work aims to change that by allowing users to build composite visualizations directly in immersive environments, such as virtual reality. This approach could lead to a more engaging and intuitive experience while helping users understand how different visualization views relate to each other.

#Composite Visualization

Composite visualization is the design that combines different visual representations into one coherent view. This could include graphs, charts, and other visual elements that help clarify relationships within the data. Traditionally, these visualizations are created outside the immersive environments by skilled designers and programmers. Users simply interact with pre-made visualizations rather than participating in their creation, which can feel limiting.

As data visualization becomes more critical for complex analysis, it's necessary to empower users to create their composite views. This can help them engage more deeply with the data and understand its various aspects. By enabling users to create visualizations within immersive environments, we are providing them with more freedom and flexibility than ever before.

#Development of a Design Space

To make this happen, we developed a design space where users can interact with various visualization views. This design space is based on how people naturally manipulate physical objects. In real life, we often grab, move, and assemble items in ways that make sense to us. By mimicking these actions in a virtual environment, users can more easily understand how to create and combine different visualizations.

Our goal was to create a set of interactions that allow users to combine visualizations effectively while keeping the process straightforward and satisfying. To do this, we broke down the interactions into manageable components. This makes it easier for users to engage with the visualizations and feel in control of the process.

#Types of Composite Visualization

We explore various types of composite visualizations by categorizing them based on how the different views relate to each other. Here are some of the primary categories:

1. Juxtaposed Views: These views place different visualizations side by side for easy comparison. This is often used with small multiples-several similar visualizations shown together to reveal differences easily.

2. Integrated Views: These views connect different visualizations with clear links, such as graphical lines or other indicators. This helps users see how the data points relate to one another.

3. Superimposed Views: In this kind of view, one visualization overlays another. This can help users make connections between the data in a more visually impactful way.

4. Overloaded Views: This category involves adding a new representation on top of an existing one without needing a direct link. The new visualization enhances the original by providing extra information.

5. Nested Views: In nested views, one visualization completely replaces a part of another. This allows for deeper exploration of the specific details of the data.

Understanding these types of views is crucial to designing usable interactions. Each type requires different interactions and user intentions, so we must consider how to best support these needs in immersive environments.

#Data Relationships

When creating composite visualizations, it’s essential to consider how the underlying data points connect. There are four main types of data relationships:

1. None: There is no overlap between the two sets of data; they are unrelated.

2. Item-Item: There exists a one-to-one relationship between data points in the two sets.

3. Item-Group: One data point corresponds to multiple elements in the other set, indicating a one-to-many relationship.

4. Item-Dimension: Similar to item-group, but one element in one set corresponds to multiple data points under a single attribute in the other.

Recognizing these relationships helps in understanding how to combine different visualization views effectively. The design must consider what combinations make sense based on the data connections.

#Creating Composite Visualizations

With our design space defined, we explore how users can create composite visualizations through straightforward interactions. Here are the key components of interactions that are established:

1. Interaction Target: Users can choose to interact with an entire visualization, a specific segment of it, or non-data elements like axes.

2. Target Manipulation: Users can perform basic 3D manipulations like changing position, rotation, scale, or motion. This allows for intuitive interactions similar to those they use in daily life.

3. View Relations: The design space takes into account how different views relate to each other, and users can manipulate these relationships as they construct their composite visualizations.

By focusing on these components, the interactions become more natural and user-friendly, making it easier for users to engage with the visualizations.

#Case Studies

To showcase the potential of this design space, we conducted several case studies. Each case highlights how different types of composite visualizations can be created through user interactions.

#Case 1: Juxtaposed Views

In this case, users create juxtaposed views by expanding the axes of the original visualization. For example, a user might grab the x and y axes of a scatter plot and stretch them to create multiple smaller scatter plots side by side. This allows for easy comparison and exploration of specific data ranges.

#Case 2: Integrated Views

Users can create integrated views by bringing visualizations closer together. When two views become close enough, visual connections like lines appear between them to indicate relationships. This helps users see how different sets of data correspond to one another, facilitating a clearer understanding of the information.

#Case 3: Superimposed Views

In this scenario, users overlay a scatter plot on a map. By lifting one visualization over the other, they can create a superimposed view that visually integrates both representations. This can help users analyze spatial data alongside other metrics, making the data more relatable and impactful.

#Case 4: Overloaded Views

Users create overloaded views by separating two axes in a parallel coordinates plot. Once the axes are apart, a scatter plot appears to provide additional context to the data. This allows for a more comprehensive analysis of the relationships between the attributes represented in the original plot.

#Case 5: Nested Views

In this case, users can nest a bar chart within a graph node. By adjusting the size of the chart and placing it strategically within a visual component of the graph, users create nested views. This format allows for a detailed examination of various attributes within the graph without losing the overall context.

#User Study

To assess the effectiveness of our interaction design, we conducted a user study. The goal was to evaluate how easily participants could create composite visualizations and what experiences they had while interacting with the system.

#Experimental Setup

The study involved 16 participants, each of whom used a virtual reality headset to engage with the visualizations. They were guided through various interactions and provided feedback on the usability and overall experience.

#Usability Findings

Participants generally found the interactions easy to learn and remember. Many noted that the design felt intuitive and aligned with their mental models of how composite visualizations should work. However, some concerns arose regarding the potential for unintended merging of views due to overlapping interaction commands.

#User Experience Insights

Participants appreciated the natural feel of the interactions, likening them to real-life movements. They noted that the experience felt engaging and enjoyable. The ability to create and decompose views freely enhanced their overall analysis process, allowing them to validate hypotheses and explore data in-depth.

#Conclusion

This work presents an innovative approach to creating composite visualizations in immersive environments. By empowering users to engage in the creation process through natural interactions, we aim to enrich their understanding of complex data relationships.

The findings of our study highlight the potential for improving user experiences with data visualization by providing tools that align with how people interact with the physical world. As immersive environments continue to evolve, the ability to easily create and manipulate visualizations will play a crucial role in enhancing data analysis and communication.

Future work should explore additional types of composite visualizations and the relationships between different data facets. By continuously developing this design space, we can further refine the ways users create and interpret data in immersive environments.