Historically, there have been large disparities in the degree to which different communities have access to resources and representation within society. With the increased availability of the internet and the growth of user-generated content platforms like Twitter and Wikipedia, there are opportunities to alleviate some these long-standing barriers to access and representation. However, there is growing evidence that many of these technologies may instead be reinforcing some of these long-standing disparities. In the first part of this dissertation, we examine how different segments of the population are represented in social media and peer production, with a particular focus on the urban-rural divide. We demonstrate that it is important to go beyond surveys of participation rates, that online representation must be evaluated in the context of different consumers of online content. Across three studies, we find that even with proportionate participation in rural areas, disparities in online representation can still remain in the quality of content viewed by users, robustness of conclusions in computational social science about these areas, and precision of algorithms that are trained from this online data. In the second part of this dissertation, we focus on the domain of geographic algorithms, evaluating how biases arise in vehicle routing, place recommendation, and geographic representation learning. We develop a framework for choosing geographic hyperparameters that affect the performance of these technologies. We provide methods for evaluating the fairness of these technologies with regards to these geographic hyperparameters. Across these studies, we find a complicated relationship between choices made in designing the algorithms underlying these technologies and the impact of these algorithms on communities. I conclude with an overview of best practices for working with geographic data and human-centered algorithms with the goal of developing technologies that are more equitable and more readily evaluated for disparities in their impact on different communities.

Creator

• Johnson, Isaac

DOI

• https://doi.org/10.21985/n2-zarw-ze54

Subject

• Computer science

Language

• en

Alternate Identifier

• etdadmin_upload_640270
• http://dissertations.umi.com/northwestern:14540

Keyword

• Algorithmic Fairness
• Structural Inequalities
• Online Representation
• Geographic Technologies

Date created

• 2019-01-01

Resource type

• Dissertation

Rights statement

• In Copyright