TourSafe — Tourism Forecasting and Recommendation (ML)

TourSafe was my undergraduate dissertation project — a machine learning system for tourism forecasting and personalised recommendation, built on content-based and collaborative filtering with time-series forecasting layered on top. Implemented in Python and deployed via a React/Python web stack, the system achieved a 20% improvement in forecast accuracy over baseline. Looking back, TourSafe was also a spatial and information problem: tourist flows encode geographic and behavioural patterns that purely temporal models cannot fully capture, and the recommendation task required thinking seriously about how information about places gets organised, filtered, and surfaced to different users.