AGA Datathon 2026
We built AuditsMadeSimple as a launchpad for understanding federal spending and audit data.
"Create a supplemental resource promoting literacy in audit terminology and data to make government engagement accessible for the general public."
We connected three major datasets (USAspending, FAC, and SAM) using the Unique Entity Identifier (UEI) to reveal gaps between funding and exclusion status.
Using Tableau, we created exploratory dashboards to highlight patterns in funding concentration (Top 10 States) and audit finding signals.
We trained a HistGradientBoosting classifier to predict next-year audit findings based on award complexity and prior history (ROC-AUC: 0.77).
We built a responsive, accessible web platform using HTML/CSS and deployed it on AWS to make these insights public.