tessakremser-lgtm/ai-social-good-project

This README describes an AI system to help manage food-waste reports and questions. It ingests unstructured text or images (complaints, photos of dumpsters, product labels), uses NLP to extract structured fields (location, waste type, urgency, responsible department, entity type), and either forwards validated complaints to the appropriate municipal department or answers consumer questions about disposal, edibility, and donation. Intended users include residents reporting waste (including low-income households and students), municipal staff who receive cases, and program managers. Key features: image/text classification, schema-based extraction, automated routing, outputs with complaint summaries or item-specific disposal/expiration advice, human-in-the-loop verification, and explicit handling of uncertainty and failure cases (e.g., unknown accused entity). The design favors accuracy and accountability over pure throughput when data are incomplete.