CAREER: Understanding and addressing geographic inequalities in location-aware technologies

Project: Research project

Project Details


Location-aware technologies have become pervasive. We now use location-aware peer economy platforms like Uber and Lyft to get from one place to another. Yelp and Foursquare help us choose where to eat, shop, and have fun. People and many prominent algorithms (e.g. the Google Knowledge Graph) learn about places both foreign and local through geotagged photos, geographic
Wikipedia articles, geotagged tweets and other types of volunteered geographic information. Social scientists are using geotagged social media for studies of a character and extent that would have been impossible ten years ago. The list of location-aware technologies that mediate our interactions with the world around us and our understanding of our world is truly massive, and it is expanding at a rapid rate.

However, evidence has begun to emerge that location-aware technologies are exacerbating the many demographic barriers that exist across our landscape. For instance, preliminary findings suggest that Uber, TaskRabbit and other location-aware peer economy platforms provide poor neighborhoods with worse service at higher prices (if they even provide service at all). Similarly,
there is more and more data in support of the hypothesis that key location-aware algorithms(e.g. location-aware recommender systems and local search systems) are more accurate in areas with certain demographic profiles (e.g. richer areas) than in areas with different profiles. Along the same lines, we and others have established that critical studies that use geographically-referenced
social media are likely systematically undercounting disadvantaged populations.

The goal of this proposal is to reverse this trend. If successful, this proposal will help ensure that the benefits of location-aware technologies are shared widely, rather being than limited to areas where people of specific demographic profiles live. The space of location-aware technologies is large, and we will focus on two families of high-impact technologies: volunteered geographic information-based technologies and location-aware peer economy technologies. Studying
these technologies will enable us to ask important questions ranging from "Are location-aware recommender systems (e.g. Foursquare, Yelp) less effective in neighborhoods with certain demographic characteristics, and if so, can we address this problem?" to "How can we increase the numberof Uber drivers from poor areas?"

This proposal also describes a series of teaching activities, all of which are tightly integrated with the proposed research. The main platform for these activities will be future offerings of PI Hecht’s massive open online course (MOOC) on spatial computing. The first offering of the MOOC had over 21,000 participants and more than 1,200 students completed the course (a high ratio for MOOCs). The MOOC’s gender ratio was 40% more female than that of majors in University of Minnesota’s computer science department, and over 30% of MOOC students came
from developing countries.
Effective start/end date8/1/162/28/21


  • National Science Foundation (IIS-1707296 003)

Fingerprint Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.