Anytime, Anywhere vs. Right Here, Right Now
Rethinking Situated Analytics
Situated analytics offers significant promise and exciting scientific challenges in data visualization, providing novel ways to integrate data analysis with physical environments. The idea of embedding data directly into our surroundings, allowing us to interact with information in the context of the real world, is compelling. Imagine walking through a factory and seeing production metrics hovering over each machine, or exploring a city with historical data overlaid on buildings and landmarks. These scenarios paint a picture of a future where data is seamlessly integrated into our physical reality.
However, despite growing availability of affordable AR/XR hardware, widespread adoption of situated analytics and situated visualization has been slow. As we stand on the cusp of this potential revolution in data interaction, I argue that this stems primarily from the poor everyday applicability of situated analytics. While the concept is fascinating, its practical implementation faces several significant hurdles that restrict its use beyond niche scenarios. The most significant one is that there are actually very few “killer apps” for situated analytics.
The Problem with Situated Analytics
The specialized nature of situated analytics presents its first major obstacle. Many data analysis tasks don’t have a clear physical analog or don’t benefit significantly from being tied to a specific location. Consider financial data analysis, social media trend exploration, or abstract scientific concepts — these often lack a meaningful physical representation or embedding that would justify the use of situated analytics. Many datasets are complex, multidimensional, and lack an intuitive physical representation. Forcing such data into a physical context might actually hinder rather than help understanding, adding an unnecessary layer of complexity to the analysis process.
Hardware constraints further compound this issue. While augmented reality (AR) technology is advancing rapidly, we’re still far from having ubiquitous, unobtrusive AR devices that can seamlessly blend digital information with our physical world. The current generation of AR headsets and mobile AR applications, while impressive, still present usability challenges that limit their practical, everyday use.
Obviously there is no disputing that there exists cases where situated analytics is beneficial. In some cases, a situated approach can be useful, even elegant; see the example images below taken from Willett et al.’s paper on “Embedded Data Representations” (IEEE TVCG 2017). However, my argument is that while these settings are compelling and offer many interesting scientific challenges, they are also exceedingly rare.
The Case for Ubiquitous Analytics
As an alternative to the constrained vision of situated analytics, I propose a ubiquitous analytics approach encompassing a broader spectrum of data engagement methods across various mobile devices, novel computer hardware, and more general settings. The key principle here is flexibility — enabling users to interact with data anytime, anywhere, and on any device.
Ubiquitous analytics doesn’t reject the potential benefits of situated analytics but rather positions it as a valuable subset within a larger framework. This broader approach recognizes that the most effective data engagement method may vary depending on the user, the task, and the context. Sometimes, this might indeed involve AR-based situated visualization. Other times, it might be a traditional desktop application, a mobile dashboard, or a voice-activated AI assistant providing data insights. Put differently, sometimes that email can’t wait even if you happen to currently be wearing an AR headset.
By optimizing for common use-cases rather than specialized scenarios, we can benefit a wider range of users and applications in interactive data analysis. This approach allows us to leverage the strengths of various platforms and interaction paradigms, providing users with a seamless data analysis experience across different devices and contexts.
Comparative Analysis
To illustrate how ubiquitous analytics encompasses and extends beyond situated analytics, let’s compare various analytical approaches (reproduced and improved from a recent situated analytics survey paper by Shin et al. in IEE TVCG 2023):
Ubiquitous analytics incorporates the strengths of other approaches while avoiding their limitations. It allows for data visualization and analysis across any platform, with or without location specificity, depending on the needs of the task at hand.
The Path Forward
While situated analytics will undoubtedly play a role in the future of data interaction, I believe our primary focus should be on developing ubiquitous analytics solutions that work in an anytime, anywhere manner. This approach doesn’t discard the unique opportunities that situated analytics presents but rather frames them within a more flexible and widely applicable paradigm.
By prioritizing ubiquitous analytics, we can
- Develop more versatile tools that adapt to users’ needs and contexts;
- Leverage the strengths of each specific device and platform;
- Ensure that data insights are accessible regardless of physical location; and
- Create a seamless continuum of data interaction experiences, from mobile to desktop to immersive environments.
As we move forward, we should remember that the goal of any analytical tool is “insights, not numbers” (freely interpreted from Richard Hamming). Whether these insights come from a hologram hovering over a physical object or a traditional screen shouldn’t matter as much as their accuracy, timeliness, and relevance. The choice of device, platform, and visual representation should not be a horse race, but rather a concerted effort to find the best solution for each unique analytical challenge.
While situated analytics offers exciting possibilities, the path to truly transformative data interaction lies in embracing a more ubiquitous approach. By focusing on anytime, anywhere analytics, we can create a future where data insights are seamlessly integrated into our daily lives, enhancing our decision-making capabilities across all contexts and environments. This inclusive vision doesn’t abandon the promise of situated analytics but rather places it within a broader, more flexible framework that can adapt to the diverse and evolving needs of data users everywhere.
