Digital Tourism From Tripadvisor Circulation Graphs to Behavior Extraction and Characterization [ADBIS’26 Tutorial]

ADBIS tutorial - digital tourism

After 8 years in DVRC lab at ESILV, joining the digital group (as new team leader) and now the Digital Tourism Working Group in the 2EMARK Axis, I’m pleased to present our joint work on Digital Tourism. In collaboration with Sébastien Jacquot at EIREST Lab at Panthéon-Sorbonne, Métropole Européenne de Lille, and Bordeaux Métropole, we studied the impact of users’ digital footprint on tourism web platform such as Tripadvisor.
We have transformed those footprints into a graph data model dedicated to traffic modelisation in Neo4j. From this, we studied tourists’ behavior in numerous manners. This tutorial at #ADBIS’26 conference (Orléans) resumes 8 years of research on the extraction of knowledge from this specific graph.

Digital Tourism has emerged as a powerful research field for understanding how tourists move, experience destinations, and shape urban dynamics through the massive digital traces they leave on online platforms such as TripAdvisor. This presentation summarizes a set of approaches that model tourist mobility using graph-based representations, particularly circulation graphs stored and analyzed in Neo4j. By building aggregating time series in a graph at multi-scale and various time frames, these methods enable both quantitative and qualitative analyses of mobility and provide key indicators based on graph advanced algorithms. Beyond mobility measurement, new approaches combining topology and geodesic measures help characterize propagation effects across territories while optimizing computations on large datasets. Advanced mining techniques—ranging from centrality normalization, sequential pattern extraction to frequent graph mining and distributing frameworks—allow the discovery of recurrent movement structures, seasonal trends, hidden patterns and early signals of latent overtourism. The talk also highlights briefly how sentiment analysis and graph models can capture evolving perceptions of authenticity in online reviews before, during, and after visiting a country of origin. Overall, the presented work demonstrates how scalable graph analytics and data mining can support decision-making in tourism marketing, urban planning, and policy design.
[https://link.springer.com/chapter/10.1007/978-3-032-39820-8_29]

A big thank you to all my colleagues, students, and collaborators with whom I have worked on this topic, an interdisciplinary field at the intersection of computer science, geography, and sociology.
Here are the slides I am presenting today.

Laisser un commentaire