Natural Language-Powered Framework for Automated Traffic Impact Assessment and Report Generation

Traditional traffic impact assessment (TIA) relies on specialized software and manual processes, creating efficiency and expertise barriers. This paper presents a natural language processing (NLP) framework that automates this workflow. Its core is a natural language understanding (NLU) engine that interprets intervention descriptions (e.g., road closures) and converts them into parameters for an open-source traffic assignment model. An integrated report generation module then produces structured, multi-format reports (PDF/Word/Web) containing key metrics, visualizations, and management suggestions. Validated on the Shenzhen road network, the framework achieves 92% parsing accuracy and reduces analysis time from hours to under 5 min. Case studies confirm its effectiveness in providing rapid, actionable insights for urban traffic management, significantly lowering technical barriers for non-expert users.