Individualized tourism itinerary planning and route optimization method integrating spatio-temporal big data

This article focuses on the shortcomings of the current travel recommendation system path planning, and puts forward individualized travel itinerary planning and path optimization methods that integrate spatio-temporal big data. By constructing user interest model and environmental state model, ant colony optimization (ACO) algorithm and graph attention network (GAT) are combined to achieve one-stop service from scenic spot recommendation to route generation. The system uses social media, traffic flow, weather changes and other multi-source data to dynamically adjust the path planning strategy. The experiment is based on a real data set of a first-tier tourist city. This data set covers 50 scenic spots and 3,000 user trips to recommend the assessment of indicators such as precision, path length and user satisfaction. The results show that the F1 value of this method is 0.756, the average path length is 23.4 km, and the overall user satisfaction is 4.5 points, all of which exceed the traditional methods. The path planning method combining deep learning and intelligent optimization algorithm can effectively improve the personalization and intelligence of tourism services.