Student Projects

Carissa Kieger '24 explored the flaws within YouTube’s algorithm and recommendation system, what actions YouTube has taken to address them and where improvements still need to be made. 

Tiana Pham '24 used Tweepy to delve into the distinct ways Twitter discussed abortion in 2021, analyzing different terminologies based on key search terms. In September 2021, Texas passed Senate Bill 8, also known as the Heartbeat Bill, that allows people to report anyone they know who is attempting to get an abortion after six weeks of pregnancy. The abortion discussions on Twitter surrounding this bill range from distant to charged depending on with which part of Twitter a person happens to interact. 

Through meta-analysis, Priya Melonio '22 explored TikTok’s non-inclusive beauty algorithms, the reason they are used and how we can improve or discourage use. With a similar dataset to TikTok's algorithm known as SCUT-FBP5500, Priya analyzed how it perpetuates toxic beauty standards that are based on a far too simple conception of what is considered beautiful. 

Freddy Pfeiffer '23 research the theory that, ultimately, cancellations resulting from so-called "Cancel Culture" are just blips and not a productive force that "speaks truth to power." Not even the largest online mob will ever be as powerful as those who control these tech platforms. 

Through data analysis, Sophie Clipson '24 explored why books become blockbusters. The Flesch-Kincaid Grade Level Test on the bestselling novels "Fifty Shades of Grey" and "The DaVinci Code" generated a "readability arc" for each book. These arcs demonstrated the importance of authorial choices in reading comprehension and if readability arcs can serve as another pattern or criteria for a blockbuster novel.

Madeline Vonk '23 used sentiment and readerly analyses to identify the emotional peaks and valleys of the plot of "Slaughterhouse-Five" to determine how the novel's themes affect its structure. Madeline explored if, when graphed, the novel has peaks and valleys that reflect the cyclical theme of renewal. Additionally, sentiment analysis may help identify which aspects of the novel are real and which are manifestations of Billy Pilgrim's PTSD from fighting in World War II.

Kaiya Case '22 studied the application of the GPT-2 neural network to classical poetry and how the more concise, metrical form may translate to machine generation. While the technology is still far from independent, the GPT-2 neural network can be trained on an existing body of text in order to emulate the linguistic style and eventually produce new, original work. 

Sejin Kim '22 explored a categorization method that can be used by researchers to break down Holocaust-related text-based resources into their component categories so that they can be analyzed with a greater degree of precision. With taxonomy techniques and resources, historians have a much better chance at finding applicable and important resources to their research that would have otherwise been unfindable.

Kaiya Case '22 assessed how the GPT-2 neural network can be applied to classical poetry and how the concise metrical forms may translate to machine generation.

Hemmi Song '24 traced the correlation in musical emotion with the emotional response to COVID-19 based on the types of music widely listened to during the pandemic. The streaming service Spotify saw a large increase in the number of subscribers, surpassing pre-pandemic estimates. The sudden increase suggests that more people were using music as a coping mechanism. The constant varying conditions of the pandemic led to highs and lows in public sentiment over the course of the past two years and an overall decrease in the mental health of the U.S. population.

Through meta-analysis, Lindsay Pantell '23 examined the correlation of the global peaks in COVID-19 cases with the increase in quarterly revenue for streaming platforms such as Disney+, Hulu and Netflix with the goal of demonstrating the human instinct to turn to entertainment at a time of heightened uncertainty and anxiety.

Celeste Ramirez Diaz '22 researched the impact of social media on Immigration Policies Public’s response to the question of whether the administration was appropriately considering the lives, safety and protection of those in detainment facilities during the COVID-19 pandemic.

Anna Barrett '23 examined digital ID programs that use Blockchain technology to store information across multiple health and social services organizations. These programs could provide unhoused individuals with a secure and easily accessible way to store identity information and health records, ostensibly breaking the cycle of homelessness.

Lara Knopf '23 looked into climate "doomism" and activism on Twitter. With the instant accessibility of media, the quick and widespread knowledge of environmental catastrophes can make the current climate crisis seem hopeless. Through the use of Tweepy, Lara analyzed tweets surrounding the subject of climate change to see the proportion of climate "doomists" versus "activists" to learn how to proceed in the climate crisis.

Using Python, Raul Romero '22 collected 316,141 tweets from 90 political leaders of the Venezuelan opposition to illustrate the changing sentiment towards Guaido and the interim government in Venezuela in 2019-2021. The project illustrates Guaido’s popularity over time using retweet count and sentiment analysis since he was sworn as interim president in January 2019 and the division of the Venezuelan opposition as well. 

Thaise Sudano '22 examined whether the current discussion around Dobbs v. Jackson Women’s Health Organization is in line with public opinion.

Through data analysis, Jill Noorily '24 explored the intersection of social media and politics, providing insights into the inner workings of social media and voter fraud. Focusing on election fraud in the 2020 presidential election, Jill collected 36,000 tweets and used topic modeling to analyze them to further understand how social media, specifically Twitter, interacts with individual views of U.S. democracy and the electoral process.

Jake Davidson '23 assessed the outlook for the Democratic Party by examining changes in voting behavior from 2016 to 2020 in five key counties: Milwaukee (WI), Maricopa (AZ), Wayne (MI), Philadelphia (PA), and Fulton (GA), all in states President Biden flipped from President Trump. Jake analyzed changes in voter turnout, votes for each party, margins of victory and percentage point victory to help quantify any liberal shift.

Sara Campagna '22 examined what is considered "real" country music. What truly is the difference between traditional country music and newer country music or is there is a difference at all?

Ani A. Parnagian '23 explored the importance of AI in the art world and its applications to discover ghost paintings. The technology utilizes a very human way of looking at negative space, and combines it with the mathematical precision and numerous style images that only a computer can process.