Abstract
My technical project and STS research paper share a theme which investigates how the choices embedded in software platforms affect user autonomy. My capstone explores how software can expand user agency by revealing information that certain platforms withhold. My STS paper examines a case in which software was used to manipulate users, which actively undermined their agency. Both projects show that software capabilities are versatile in that the same features that empower users can just as easily be used to harm them. My capstone, “Enhancing Housing Search Efficiency: Integrating Public Transit Analytics into Real Estate Listings,” addresses a practical gap in the NYC apartment rental market. StreetEasy, the platform most NYC renters rely on, provides no transit commute estimates, forcing users to cross reference every listing against Google Maps. This is a process that can take thirty to sixty seconds per listing, which can amount to hours over a full apartment search. I built Pace, a Chrome extension that injects transit commute times into StreetEasy listing pages using JavaScript DOM manipulation and the Geoapify API. Users set a workplace destination and Pace displays commute times, keeping renters informed without needing to leave the page. My STS paper, “A Kantian Analysis of Facebook’s 2012 Emotional Contagion Experiment,” examines platform design used for the opposite purpose. In January 2012, Facebook secretly modified the News Feed algorithm for nearly 700,000 users to filter the emotional tone of the content they saw, without their knowledge, in order to test whether emotions can be spread digitally. Most critics treated this as a procedural failure of missing informed consent, whereas my paper argues that view is insufficient because it treats the algorithm as a neutral tool rather than a tool of psychological control. Applying Kant’s second formulation of the categorical imperative, which states that one must treat humanity always as an end and never merely as a means, I argue the deeper moral failure was architectural. The engineers deliberately bypassed users’ rational agency in a way no consent process could have fixed. Building Pace gave me direct experience with the technical decisions that determine what users see on a platform, which helped make the mechanics of the Facebook experiment more clear. When I analyzed how Facebook’s engineers filtered emotional content from the News Feed, I was reasoning about decisions I had faced myself. Researching the Emotional Contagion experiment in turn led me to think about Pace’s own design more carefully, such as what data it collects, how preferences are stored, and whether the tool actually serves users or merely processes them.
