Projects
Illegal Dumping
in Philadelphia
Manuscript in Preparation
In partnership with the City of Philadelphia's Department of Parks & Recreation and the Water Center at Penn, this project applies spatio-temporal Bayesian models to over 100,000 city 311 service reports to identify persistent illegal dumping hotspots across Philadelphia. By jointly modeling spatial structure, temporal dynamics, and socioeconomic factors, the analysis uncovers where and when dumping concentrations form and persist — with particular attention to community parks and green spaces in socio-economically vulnerable neighborhoods. Findings inform more targeted intervention strategies for city agencies.
DataPoints
Actively MaintainedData Points showcases the work of Penn's data science community through concise, engaging articles. Each post takes a complex idea and transforms it into accessible insights in creative and compelling ways — whether through a high-level walkthrough of a key figure, or an interactive, explorable explanation.
Built entirely with HTML, CSS, and JavaScript from a self-designed Figma prototype. The visual identity takes Penn's institutional color palette as a starting point and pushes it toward something more expressive — trading the standard deep red and navy for light pink and mint blue to keep the reading experience open and inviting. Every layout decision, from the article grid to the post template, was designed from scratch. Each post's cover image was also individually designed and produced.
AI Resource Page
Actively MaintainedA curated resource hub for the Penn Arts & Sciences community, bringing together practical LLM tools and guidance for both general academic use and research-specific workflows. Alongside the resource listings, the site includes a dedicated section on responsible AI use — covering known limitations, common misuse patterns, and considerations for scholarly and research integrity.
This was a first experiment in vibe-coding with Claude CLI, with site architecture, content decisions, and all editorial choices designed, reviewed, and refined by me and colleagues at DDDI.
Broadway Musical
Database
Ongoing
A relational database of Broadway musicals built from data web-scraped from the Internet Broadway Database (IBDB), paired with an interactive ArcGIS StoryMap that explores the data spatially and narratively.
Note: the underlying scraped dataset cannot be shared publicly due to IBDB's data policy. The repository contains the scraping and processing code; the StoryMap below presents the resulting analysis. Work is ongoing but currently paused — all output shown reflects progress as of 2023.