Projects
PennOS
April - May 2024
In this project, we successfully built a single-core operating system, with a FAT-based filesystem, a kernel, and a scheduler that correctly decides which processes to run. We have preserved the necessary abstractions between kernel, system, and user land. We have implemented a number of builtin functions that can be run from our shell and interact with the filesystem. We have tested the functionality of the entire system, including the correct CPU utilization and memory leaks using Valgrind.
Partners for Justice
Feb - May 2024
In collaboration with Partners for Justice, a national nonprofit organization dedicated to transforming the legal system, I developed a tool that automates data analysis for advocate databases stored in Airtable. This tool addresses the challenges of manual data analysis and repetitive tasks across multiple databases as the number of advocate offices grows. Using the Airtable API, the tool pulls data from specified databases and performs queries to calculate key insights, such as the number of cases opened and the comparison between entered and provided service goals. Users can specify date ranges for targeted analysis, reducing a process that once took days to just minutes. This tool significantly enhances efficiency and accuracy in data reporting and analysis, supporting Partners for Justice's mission to save clients from incarceration and connect them with vital social services.
AirBnB Price Prediction Machine Learning Model
April - May 2023
In this project, we aimed to better understand the drivers of price for various Airbnb listings in New York City in 2019. We selected three datasets to achieve this goal. The first dataset, airbnb_ddf, contains information on price, geographical location, neighborhood, room type, reviews, and other features of Airbnb listings. The second dataset, crime_ddf, provides detailed records of various crimes committed across New York City, including their classification (felony, misdemeanor, or violation) and geographical location. The third dataset, income_df, offers data on the median household income for each of 200+ neighborhoods, adding complexity to our analysis. By aggregating these datasets and merging them, we developed a comprehensive understanding of both intrinsic factors (e.g., room type, reviews) and external factors (e.g., crime levels, income) that influence Airbnb listing prices.
FIFA Database Application
April - May 2023
In this project, we successfully developed a comprehensive web application for analyzing and comparing football players and teams across various leagues and seasons. The application allows users to track the yearly progression of individual attributes, such as overall rating, shooting, and dribbling, and compare multiple players and teams across these attributes. Users can filter and search for specific players and teams, with visualization features like graphs and charts to enhance comparisons.
CommonCents Web Application
Sep - Dec 2021
In collaboration with Penn Spark, we developed CommonCents, a gamified edtech platform designed to empower students with the knowledge to manage their money and achieve financial independence. This platform aims to make personal finance learning innovative, interesting, and inclusive for a diverse audience, ranging from high school students to undergraduates. As a team of four developers, we drove the application from development to deployment. CommonCents now attracts over 2000 monthly visits across 8 campuses and serves 1600 students.
