This report links APD property crime incidents (2018–2025) to tax parcels and Census tracts to show where crime clusters, what kinds of land uses are most exposed, and how those patterns persist over time. Using spatial statistics (Moran’s I, LISA) and tract-level regression, we highlight stable hot spots along major commercial corridors and the Atlanta BeltLine and discuss implications for planning, lighting, parking security, and neighborhood stability.

Using Reddit threads that mention “stealing” in Atlanta-related subreddits (2018–2025), this project tracks how people talk about property crime over time, by day and time of posting. I combine word networks, word clouds, and two sentiment models (dictionary and BERT) to compare narratives about theft, from car break-ins and apartment garages to lighter, joking uses of “stealing.”

What makes one neighborhood pleasant to walk through and another uncomfortable? This project explores that question using Google Street View and computer vision. Focusing on two Atlanta Census Tracts, one walkable and one unwalkable, I extracted visual features such as greenery, sidewalk presence, and street enclosure. This project shows how modern visual data tools can help us “see” the urban environment and better understand the elements that shape walkability.

A simple spatial analysis of restaurants in Duluth, Ga

Collecting POI data using Google Places API