Given emerging diverse spatio temporal network (STN) datasets, e.g., GPS tracks, temporally detailed roadmaps and traffic signal data, the aim is to develop a logical data-model which achieves a seamless integration of these datasets for diverse use-cases (queries) and supports efficient algorithms. This problem is important for travel itinerary comparison and navigation applications. However, this is challenging due to the conflicting requirements of expressive power and computational efficiency as well as the need to support ever more diverse STN datasets, which now record non-decomposable properties of n-ary relations. Examples include travel-time and fuel-use during a journey on a route with a sequence of coordinated traffic signals and turn delays. Current data models for STN datasets are limited to representing properties of only binary relations, e.g., distance on individual road segments. In contrast, the proposed logical data-model, Lagrangian Xgraphs can express properties of both binary and n-ary relations. Our initial study shows that Lagrangian Xgraphs are more convenient for representing diverse STN datasets and comparing candidate travel itineraries.