Calculating Risk: A Financial Application of a Cloud-Distributed Platform; A TikTok Case Study: The Ethics and Externalities of Data-Gathering Technologies

Ham, David, School of Engineering and Applied Science, University of Virginia
Earle, Joshua, Engineering and Society, University of Virginia
Morrison, Briana, Computer Science, University of Virginia

Computer science is often viewed with an extreme polarity. Hype for rapid change in industry and everyday life is counterbalanced by a growing distrust of all things digital. This thesis portfolio feeds into both sides of this dichotomy within computer science. In my technical project, I discuss a personal experience which shed light on how programs are changing modern finance, showing the intersection between UVA computer science education and industry. In the STS research paper, I perform a case study in which I investigate the effects of diminished internet privacy today in social media platforms. Both sections are totally distinct, yet they are two topics that capture the love-hate relationship a lot of people have towards programs.

The STS portion of this portfolio examines the social media giant TikTok and the risks inherent in their aggressive collection of user data. More specifically, the paper delves into how TikTok’s collection of user data and algorithmic uses of this data work to eliminate human choice and intellectual freedom, resulting in polarized relations on both a person-to-person and global scale. Starting off with a historical background, the paper explores the company’s scandals and its developing relationship with the Chinese and United States governments. Since TikTok had gone international, their relations with the US progressively worsened, also breeding distrust between the US and China. The paper goes on to look into the kinds of data TikTok collects and their most immediate uses of this data, primarily looking into the company’s privacy policy and studies on their “For You” page’s video recommendations. This section’s findings reveal that TikTok’s highly individualized data on each user enables the platform to hijack the user’s attention with the recommended content, eliminating the need for the user to choose what content they view. Given that TikTok’s userbase is largely children, this gives a computer algorithm a concerning amount of power over the ideologies and mental beliefs a person may develop from a young age. The paper’s final section explores the broader societal consequences resulting from a platform’s extreme personalization, looking into the Cambridge Analytica scandal and an analysis on highly segmented advertisement and political discourse. When everybody is fed a different truth, people’s ability to participate in healthy discussion of modern events is minimized, harming the democratic process as a whole.

The technical portion of this portfolio expands on a summer internship experience I had going into my fourth year. During my internship with a consulting firm, I worked on an engagement with a major investment bank, and my team was tasked with implementing a platform for pricing trades and risk. The investment bank had already built out its own platform using proprietary software, but my team of consultants sought to create a more efficient solution through third-party cloud-computing resources. Normally, these financial calculations involve numerous, time-consuming simulations and other computations for a batch of trades; however, by dividing up the task into subtasks and distributing them onto a grid of servers, they could be completed in parallel, overnight. My focus during the summer was on the “middleware” of the platform, which was the code responsible for packaging and sequencing tasks requested from the frontend and placing them on a queue to be sent to the backend for computation.

It was my task to review and present code improvements for the middleware to the broader team and implement several new features essential to the middleware’s functioning. I added two major features to the middleware and also presented to the broader team the inner workings of the platform’s dependency graph, a data structure that eliminates redundant tasks and ultimately places the correct sequence of jobs to be performed onto the queue to the backend grid. As a result, the team gained a more unified and cohesive understanding of the middleware, whose code the other workstreams of the project never directly see. By the time my internship concluded, only one of the six steps for pricing a trade had been implemented. The team must continue to implement the remaining steps for a fully-functioning platform.

BS (Bachelor of Science)
TikTok, Data, Internet Privacy, Case Study, Algorithm, Financial Risk, Concurrency, Manager Worker, Dependency Graph, Social Media Platform, Privacy Policy, Addiction, For You Page, AWS, Cloud Computing, Data gathering
All rights reserved (no additional license for public reuse)
Issued Date: