Immersive Methods for Exploratory Analysis

1. Introduction In a sense virtual reality (VR) is in a situation similar to the situation that artificial intelligence was some two decades ago. That is, it is a technology that has received much popular attention and for which the publicity received has far overshadowed the reality. The acid test of an artificial intelligence system was that the user of the AI system would be unable to distinguish between a real human and the AI system. An ideal virtual reality system would allow the user to put on some hardly noticeable equipment and enter an artificially generated, dynamically adjustable world which, as in the case of AI, was essentially indistinguishable from reality. Of course, neither AI nor VR are capable of satisfying these exceptionally difficult tests. The ideal of VR is far removed from the current practice. Our interests in VR date from the talk of Interface Symposium. In the intervening six years, we have accumulated some substantial amount of equipment and some degree of experience with the application of these technologies to the statistical setting. In this paper we will describe some of the hardware and its limitations and relate some of our experiences with this technology. Bear in mind that it is exceptionally difficult to convey to the reader the visual