The problem: There are limits to the algorithmic analysis of extremely large data sets (big data). One solution is to aggregate and reduce complexity and use visual representations (Visual Data Mining). However, this reductive approach is difficult for the exploration of more subtle and emergent patterns occurring between multiple variables within the data.

Our solution: Research from psychology suggests that virtual environments dense with information are easier to process than simplified graphic encoding – if there is alignment of the data representation with human ecological perception of natural environments. We leverage the capacity of the human senses (visual, spatial, aural and kinaesthetic) to enable the immersive analytics of big data with virtual reality.

Current state of research: The implementation of a prototype immersive analytics application based on a terrain metaphor, has been developed. The application employs a two-stage interface: in desktop mode the user experiments with assigning attributes; in VR mode the user can explore the data landscape, identify and tag patterns, anomalies and connections, with up to 7 million data values being represented in one sector. A data streaming module pre-loads data into adjacent sectors to enable potential application to big data sets.

Next stage that requires funding: Access to data sets and end users is required to evaluate the effectiveness of this novel approach. The aim is to enable end users (domain experts who not necessarily familiar with data analytics) to experiment with immersive analytics, as a way to identify emergent trends, anomalies and connections within big data.

Potential industry partners: Over 80% of big data is spatial (geo-tagged coordinates) so potential industry partners are wide, including built environment, health, economics and transport logistics.