Research at AURA Lab.

People show up in virtual environments differently than they do in messaging or video. A VR avatar carries posture, gaze, and proximity in ways a profile photo can’t. AURA Lab studies the affordances of these spaces — what they make possible, what they exclude, who they invite in — using corpus-scale platform data and focused qualitative studies.

Livestreaming compresses a parasocial relationship into a real-time channel: the streamer is performing for an audience that is also performing for itself in chat. The lab works with a 22-million-row corpus of chat-and-stream data, plus follow-up restricted-access datasets, to study how identity tagging, harassment, and parasocial communication play out in this mixed sync/async environment.

Computational methods only license the claims their assumptions support. The lab builds and uses corpus-scale tools (structural topic modeling, multi-instrument sentiment ensembles, network and community-detection methods, increasingly local LLMs) on real platform data, and ships the methodology alongside the findings. Open-source outputs include the TASS sentiment ensemble, the v2v R teaching package, and a growing set of reproducible Quarto workflows.

The Beyond Meet Space initiative — originating in a closed NSF FW-HTF-R grant — examines how virtual meetings work as communication environments rather than as broken substitutes for in-person interaction. Active threads include selective muting as an inclusion tool, impression-management features, and longitudinal studies of VR classrooms.