Abstract
Both projects in this portfolio are united by a single question: what happens when data becomes a tool of control rather than a resource for fairness? The technical capstone, Hoos Helping, confronts this question in the context of gig labor by building a community service platform that makes payment and job-matching information fully visible to all users. The STS research paper, "The Metrics of Power: How Streaming Data Opacity Reshaped Creative Labor in Hollywood," confronts the same question in a different arena, analyzing how Netflix's deliberate withholding of viewership data undermines the bargaining power and creative agency of the writers, showrunners, and artists who produce its content. Though one project is a working web application and the other a qualitative discourse analysis, both are fundamentally concerned with who gets to see the data, who gets to act on it, and who is facing the consequences when the numbers are hidden.
Hoos Helping is a full-stack web application designed to connect University of Virginia students and Charlottesville-area residents through a transparent, community-oriented gig-work platform. The project was motivated by a gap in the local market. There was no dedicated service platform that existed for short-term, person-to-person tasks like furniture moving, pet sitting, or item assembly. Moreover, there was a shared frustration with how existing platforms such as TaskRabbit and Uber operate. Those platforms conceal their algorithmic job-matching logic and withhold compensation details until after workers commit to a task, leaving them with little leverage and no clarity. Hoos Helping addresses these design failures directly. It removes algorithmic matching altogether, allowing users to browse and accept tasks on their own terms. It requires posters to display payment rates upfront, eliminating hidden fees or after-the-fact adjustments. Built with React on the front end and Django on the back end, with a PostgreSQL database and Google Authentication for verified access, the platform was developed across five two-week sprints and evaluated through usability testing with UVA students. By December, it delivered a functional application that models what an ethical, transparent gig-work infrastructure can look like at the community level.
"The Metrics of Power" argues that Netflix's cancellations of Anne with an E (2019) and Warrior Nun (2022) reveal an algorithmic culture that systematically deploys data opacity as a tool of corporate power, devaluing artistic merit in favor of engagement metrics that only the platform itself can access or interpret. Drawing on Langdon Winner's argument that technologies embed the values of their creators and Veena Dubal's concept of algorithmic wage discrimination, the paper examines Netflix's corporate statements, creator responses, and fan campaign materials through qualitative discourse analysis. The analysis finds that Netflix's cancellation language, or its deliberate absence of language, functions to make its decisions appear objective and unchallengeable. Creators like Moira Walley-Beckett, who wrote that it is "impossible to argue with words like Economics, Algorithms, Demographics," are left without meaningful recourse, because algorithmic framing converts human editorial judgment into the appearance of neutral, data-driven outcomes. The fan campaigns that followed each cancellation further illustrate this asymmetry: the #RenewAnneWithAnE petition gathered over 1.78 million signatures and received no response from Netflix. In contrast, the #SaveWarriorNun campaign achieved only a partial victory by circumventing Netflix entirely rather than changing its decision-making. The paper concludes that meaningful accountability in the streaming industry will require not just better corporate behavior, but structural access to the data on which creative decisions are made.
Working on both projects simultaneously produced insights that neither could have generated in isolation. Building Hoos Helping made abstract claims about data opacity feel concrete and immediate. Deciding exactly which pieces of information to show workers—pay rates, job descriptions, poster identities—forced engagement with the real design cost of transparency. Every feature involved a tradeoff, and that experience gave the STS analysis a more grounded appreciation for what Netflix's opacity actually conceals. When Walley-Beckett wrote that she could not argue with an algorithm, the technical project helped illuminate what that means in practice: if a platform chooses not to surface the criteria by which decisions are made, those criteria become structurally unchallengeable, not just rhetorically so. The STS research, in turn, sharpened the technical project's sense of purpose. Reading about the WGA Strike and the ways that streaming platforms use data privatization to suppress worker leverage reinforced why Hoos Helping's features, which show pay upfront and remove algorithmic matching, were not just usability improvements, but a deliberate refusal of the opacity that larger platforms treat as a competitive advantage. Together, these projects argue that the design of information systems is never neutral: it either concentrates power or distributes it, and the choice between those outcomes is one that developers and researchers alike have a responsibility to make deliberately.
