A mobile-first web app for exploring public safety data in Dorchester’s Talbot Norfolk Triangle (TNT) neighborhood through chat interaction with a large language model and a dynamic neighborhood map. Aims to help the TNT community better understand and share their experiences with public safety by combining community voices with official city data.
Built a recommendation system to guess user ratings for unseen movies. First attempted a Collaborative Filtering approach, then improved performance with an open-ended ensemble solution, motivated by the famous SVD algorithm.
Trained a binary classifier to predict age suitability of text excerpts with 70% AUROC accuracy. First attempted a Bag of Words feature representation, then improved performance with our open-ended solution.
A full-stack inventory management system for a local non-profit theatre, developed alongside a 10-person agile development team. Work was organized under JumboCode’s pro bono software development service. 