XRFlow: Visualizing Metrics and Insights for Extended Reality

Extended Reality (XR) applications combine virtual and physical elements, generating continuous data streams that enable us to analyze interactions between the users and the virtual objects. With the extensive volume of physiological and behavioral information, processing and obtaining an overview of the recorded data is often difficult and time consuming. Existing approaches typically replay recorded sessions or deploy fixed analysis pipelines. Such approaches risk overlooking time-critical or out-of-view interactions, as pipelines are only designed to detect the specific conditions they were designed to capture. This paper introduces XRFlow, an object-centric data exploration tool comprising two components: a Unity plugin that simplifies the acquisition of object- and event-based data, and an external visualization interface that renders both these data and auxiliary data streams from objects within the environment. Using a simulated scenario, we illustrate how exploratory analysis through XRFlow can uncover subtle errors and provide comprehensive overviews of XR sessions for developers and researchers.