David Noël - A semantic web based approach to study life trajectories

14:00
Monday
17
Jun
2019
  • Thesis defence
Place: 
Organized by: 
David Noël
Speaker: 
David Noël
Teams: 
Keywords: 
  • citizen as sensor
  • crowd sourcing
  • data visualisation and acquisiton
  • urban migration
  • residential trajectories

 

Jury :

M. Alain Bouju, maître de conférences HDR à l'Université de la Rochelle, rapporteur
M. Gilles Falquet, professeur associé à l'Université de Genève, rapporteur
M. Didier Josselin, directeur de recherche au CNRS, examinateur
Mme Marlène Villanova-Oliver, maître de conférences HDR à l'Université Grenoble Alpes, co-encadrante de thèse
M. Pierre Le Quéau, maître de conférences HDR à l'Université Grenoble Alpes, co-directeur de thèse
M. Jérôme Gensel, professeur à l'Université Grenoble Alpes, directeur de thèse

The notion of trajectory is the subject of many works in computer science. The life trajectory has several peculiarities which distinguish it from the trajectories usually considered in these works. It is first of all its temporal hold, which is the life, the existence of the observed subject. It is then its thematic hold, this one potentially concerning multiple aspects of the life of an object or an individual. Finally, it is the metaphorical use of the term trajectory, which refers more to the meaning of the trajectory than to the description of a simple evolution in time and space. The life trajectory is used by the expert (sociologist, urban planner...) who wishes to put in perspective the information on individuals to better understand their choices. The motivations for studying the life trajectory are depending on the application and themes considered : the relation to work and employment, family life, social life, health, residential trajectory... 

We propose a Semantic Web based approach to study life trajectories, which allows their modeling, collection and analysis. This approach is embodied by a software architecture whose components are configurable for each application case. This architecture is based on a life trajectory ontology design pattern, as well as a model of explanatory factors for life events. To operationalize the proposed modeling, we designed algorithms that allow the creation of a life trajectory ontology by exploiting the previous pattern and model. For data collection, we developed APIs to facilitate i) the construction of a model compliant data collection interface ; and ii) the insertion of the collected data into a Triple Store. Our approach allows the representation, and hence the collection and exploitation of multi-granular information, whether spatial, temporal or thematic. Finally, to allow the analysis of the trajectories, we propose generic functions, which are implemented by extending the SPARQL language. 

The methodological approach and the proposed tools are validated on a case study on residential choices of individuals in the Grenoble metropolitan area by highlighting the characteristics of their residential trajectory and the explanatory elements of it, including from their personal and professional trajectories