serenata-de-amor

Why we included this project

Serenata de Amor is a civic-tech project that applies machine learning to Brazilian public spending records, flagging expense reimbursements from congress members and municipalities that look suspicious. Its analysis engine, Rosie, pairs with Jarbas, a web interface where anyone can browse expenses and check each flagged case. The project covers the full pipeline, from scraping and cleaning open government data to training models and publishing findings, which makes it a practical reference for building oversight or transparency tools. The repository is no longer updated frequently, so treat it as a learning resource and architectural blueprint rather than something to deploy as-is. Teams working on civic data, government transparency, or public-expenditure analysis will find the approach and code worth studying.