Can You Measure How Architecture Makes You Feel?

The Bentway

Once an underutilized area under the Gardiner Expressway in Toronto, The Bentway was revitalized in 2018 as a community space for public art, events, and civic activities. As a “shared backyard” and programming space, The Bentway team was interested to know what cutting-edge sensor technology could teach their organization about the future of public space.

The collaboration with Autodesk would allow The Bentway to capture data that might help engage their local community more effectively and inform future design decisions. In turn, as a living piece of infrastructure and public park space, The Bentway was a unique site to explore the Autodesk Research team’s human centric building design hypothesis in the real world.

THE EXPERIMENT – From Steps to Stories

To understand and predict how objective features in a space connect with subjective experiences, the Autodesk Research team built an experimental participatory design platform in collaboration with The Bentway. Participants engaged in “walkshops,” in which they walked the length of The Bentway as multimodal data about their experience was recorded in various ways:

• Quantitative data: Participants wore sensorized data-capture helmets, equipped with cameras that recorded 360 video and footage of participant faces. Helmet sensors captured temperature, pressure, humidity, and luminosity. The team also tracked GPS data of participant paths.
• Qualitative data: Participants answered survey questions along their journey, for example, how lonely or social is the space? How inspired does it make you feel? They were also able to record their spoken thoughts.

THE RESULTS – Mapping Human Response to the Built Environment

To tell the story of the data, the Autodesk Research team used visualizations to convey the experiences of walkshop participants. They used two distinct datasets to analyze the relationship between spaces and human reactions: environmental features and human responses.

Evaluating the Environment

Using an object detection and scene inference algorithm, the research team extracted notable objects and the relationships between them from the 360 videos. This enabled researchers to repaint a picture of what the participants saw and correlate it with their personal experience.

Evaluating Human Response

To visualize the participant’s facial expressions while maintaining anonymity, researchers ran an algorithm to extract the facial contours and drew them as outlines overlaid on top of the video.

The research team also asked participants to evaluate their agreement with statements about the space at certain moments along the route, for example:

• It is pleasant to see this area
• I feel safe walking in this area
• I could imagine meeting with friends here
• This area is full of activity
• The air is clean in this area

Participants were also invited to record their thoughts directly into their mobile devices.