Data Science

Data Science disruption: Virtual reality

26th Sep `16, 02:06 PM in Data Science

For anyone familiar with the field of data science, it is easy to see how virtual reality can…

Philip Piletic
Philip Piletic Contributor

For anyone familiar with the field of data science, it is easy to see how virtual reality can be a natural partner for big data challenges. Oftentimes, the most difficult aspect concerning big data is not the visualization process itself per se, but rather how well scientists and other researchers are able to gain insights from the data at hand. The phenomenon of observing human-friendly visualizations of big data in well-designed models, yet not being able to make meaningful new insights, has been documented by researchers in a paper part of the Big Data Journal, which attributed the fundamental processing issue as a limitation arising from human perception itself.

Enter, Virtual Reality

It is in this specific challenge, where traditional modeling practice fails to rise to the test, that virtual reality can find its place in the future of big data visualizations. While popular virtual reality notions today focus on applications mainly in entertainment such as gaming, a rising notion has been the idea of implementing data visualizations using virtual reality technologies for science and engineering. Within this concept, using virtual reality for data visualization is a novel idea that evokes imagery currently found in science fiction films of people being able to manipulate projected data models, such as Tony Stark designing his Ironman suits or even the map of space from Prometheus.

The diverse potential of VR and AR applications

Virtual reality has made a promise, and big data practitioners are listening. Articles such as this one featuring in Data Science Central are becoming commonplace, contemplating the position and drive behind big data visualization efforts and the abilities of companies to keep up. The immersive experiences that data scientists so clearly crave have an enticing number of possible applications in all fields of science, from chemistry to engineering to astronomy. Outside of science, it is also clear that in the future, the companies that are able to optimize and restructure their processes to take into account big data insights are the ones that will prosper. This commercial early adoption drive is shown in the advent of new big data systems such as Hadoop.

Challenges and Limitations

While a popular notion, the process of creating data visualizations using virtual reality technology poses new and unique challenges for developers. Not only must they face traditional data science questions concerning how to present the actual data, they must also make it interactive enough to be able to reach viewers on a human level and inspire new research questions. While some of the technologies involved in creating these dreams of making big data “real” to users are already around today, others must be further developed in order to be of any meaningful use.

One example for which new ideas are required are optimizations concerning the storage of the data itself, for which the industry leans towards using the Cloud. However, when choosing to progress with Cloud storage, further questions arise concerning speed, security, and so on. It can be easy to see how while the general idea of big data visualizations using virtual reality may be simple, the application of it is most certainly more complicated. Even with improved storage methods, developers would still face the initial hurdle present for all data scientists: how to create a model that is as objective as possible, yet can be understood by humans. Only further explorations into this field could provide expertise on how to best solve the challenge of these three-dimensional, interactive data models.

The Early Future

Projects exist already that build upon virtual reality headsets such as the Oculus Rift, in development by Facebook, to create big data visualizations. One such project by researchers at MIT, sought to model Twitter comment data on the university campus for analysis via the Oculus Rift headset. Largely successful, this project is just one of many seeking to explore the types of interactive models that can be created using virtual reality.

MIT is not the only institute investing time into developing big data applications for virtual reality headsets. Caltech also released a study in 2014 that sought to use virtual reality as the answer for visualizing high-density astronomical data sets that normally would take months to not just peruse, but understand the data. The Caltech researchers concluded that the benefits of immersive experiences for their case provided an avenue to collaborative data visualization and exploration unfeasible for existing traditional data models. Immersion would not only provide a natural space for scientists and colleagues through a shared visual space, but also lead to better perception of the inherent datascape geometry, allowing for a more intuitive understanding of the data and even better retention of perceived relationships.

Not Just For Scientists

The expansion of VR into the market has resulted in a burgeoning platform of choices at all levels from everyday consumer to cutting-edge scientist. While the idea of truly meaningful visualizations of big data for the purposes of science may be still waiting in the near future, the devices of today are still to some level capable of demonstrating the potential of virtual reality for data science.

Out of the existing options, the Samsung Gear VR is the most consumer-friendly due to its affordable price that comes without the cost of major feature sacrifice. The Samsung Gear VR is compatible with a variety of Samsung devices, including the Galaxy Note 5, Galaxy S6/S6 Edge/S6 Edge+, Galaxy S7/S7 Edge and the Galaxy Note 7. Majority of these devices still cost quite a bit, but then again, people are more easily persuaded to buy a phone that they’ll carry with them every day then a new piece of exotic tech like Oculus. This also means that for developers not part of organizations with large enough funding to obtain more high-end devices as the Oculus Rift, there is still always the possibility of exploring big data potentials in the middle and lower reaches of the existing headset market.

Even for more everyday big data or virtual reality enthusiasts, the opportunity exists for furthering the development of virtual reality applications on big data. This accessibility to topic will ensure that will the buzz around virtual reality and big data will not only not die down, but likely only increase in interest as further headset development continues. Who knows? Perhaps in the near future we will ourselves have the ability to use virtual reality headsets for even mundane big data challenges, such as investigating how to best curate our social media profiles or monitor home resource usage. The future of meaningful big data virtual reality is coming, and it is in the technological sense, right around the corner.