EpiTrip

Project description

EpiTrip understands free sentences ("I'm leaving Rennes to go to Biarritz"), validates the travel intent, extracts cities (departure/arrival) via NER (CamemBERT, DistilBERT, FastText), then calculates an optimal route based on SNCF data and a graph algorithm like Dijkstra. The results are displayed on an interactive map with steps, durations, and a clear summary. Front: Nuxt 3, Vue 3, TypeScript, TailwindCSS. Business Backend: AdonisJS. AI Backend: Python/FastAPI (Transformers, PyTorch, scikit-learn, Pandas, Matplotlib) with optional voice recognition (SpeechRecognition). The project demonstrates a concrete application of NLP and AI to travel, accessible and educational.