• Login
  • Register

Work for a Member organization and need a Member Portal account? Register here with your official email address.

Project

Conversational Spatial Analytics in Augmented Reality

Copyright

Matti Gruener

Matti Gruener

Project Contact:
Groups

The data we collect about the physical world keeps growing. Generative models, cheaper compute, and better sensing are pushing dataset sizes up quickly, and that trend shows no sign of slowing. Bigger datasets let us draw more insight from the environments we live in, from a single building to a city to the planet. They also get harder to make sense of. A spreadsheet or a dashboard does a poor job with information that is really about places and how they relate to one another.

Take a synthetic population: thousands of simulated people moving through buildings and streets, run against different designs and policies. The data is rich, and it points toward real questions about how to live more sustainably and comfortably. It is also spatial. It reads best when you can see it sitting in the place it describes.

This project explores how to work with that kind of data inside augmented reality, using a conversational voice agent as the way in. You ask a question out loud. The agent retrieves the relevant data in near real time and renders it as a visualization anchored to the space around you.

Most data tools expect you to know the query before you start. You pick the filter, choose the chart, and read the result. Here the order is looser. You can ask for a specific value or a filtered view, and you will also be able to pose an open question and let the agent work out results and how to communicate them. The visualization is something the agent produces as part of its reply, so the explanation and the picture that supports it arrive together.

Urban and environmental data usually lives with specialists. The people most affected by it, including residents, students, and the officials who write policy, rarely get to handle it directly. Putting the data into a shared physical space, described in plain language and made visible where it belongs, gives them a way to reason about the places they care about. It can support public consultation, teaching, and the kind of evidence a planning decision should rest on.

There are many open questions here. Do people actually learn better from their data if they can interrogate it with a voice agent? Are the insights richer? Do we get to them faster? Can the agent open complex data to people who were shut out of it? Can this lead to meaningful change in the hands of legislators?