Towards Contextualized Road Safety Conditions
PI: Jacob Thebault-Spieker
Assistant Professor, The Information School
University of Wisconsin-Madison
Framework component: Interaction
Spatial technology – such as location-based games, mapping and navigation, and autonomous vehicles – increasingly rely on large-scale crowdsourced geographic information, or user generated content, as the underlying geographic data for user-facing applications, as well as the machine learning models that underpin them. Some early research in this space is identifying situations in which spatial technologies which facilitate and augment people moving through physical space are enabling or creating dangerous situations. Further, it is likely that such technologies, which commonly rely on biased crowdsourced geographic information, will reflect the geographic biases along urban-rural, socioeconomic, and demographic lines that exist in the underlying data. Therefore, this project focuses on the intersection of these two concerns with spatial technologies: disparities in the crowdsourced data that spatial technologies rely on, and the spatial technologies’ role in facilitating unsafe conditions for users. Specifically, this work will focus on the capacity for user-facing spatial technologies – relying on current, biased, crowdsourced geographic data – to support contextual road safety information. The underlying question of this proposal is: given these risks, how effectively can spatial technologies characterize road safety conditions to users?
Key Personnel
Yaxuan Yin
Graduate Student, The Information School
University of Wisconsin - Madison