Aline Menin - eSTIMe: a Visualization Framework to Assist the Analysis of Daily Mobility Data

09:30
Thursday
26
Nov
2020

 

Jury composition:

  • Thomas Devogele, professeur, Université de Tours, rapporteur
  • Sara Irina Fabrikant, professeure, Université de Zurich, rapporteuse
  • Christophe Hurter, professeur, Ecole Nationale de l'Aviation Civile, examinateur
  • Patrick Reignier, professeur, Grenoble INP - ENSIMAG, examinateur
  • Julie Vallée, DR CNRS, Laboratoire Géographie-Cités, examinatrice
  • Marco Winckler, professeur, Université Côte d'Azur, examinateur
  • Paule-Annick Davoine, professeure, Université Grenoble Alpes, directrice de thèse
  • Luciana Nedel, professeure agrégée, Université Fédérale du Rio Grande do Sul, co-directrice de thèse
  • Sonia Chardonnel, chargée de recherche CNRS, Université Grenoble Alpes, co-encadrante de thèse

The research field of urban mobility aims at the observation and design of human trips within an urban environment, which information supports decision-making and problem solving within public policies. In this context, there are many experts -- not necessarily transportation specialists -- that need to handle more or less standardized urban data to extract synthetic and easily exploitable knowledge. Hence, public transportation agencies commonly conduct trip-based surveys to collect information about day-to-day travel of the population within a particular territory (i.e. where and when we travel), resulting in large and complex datasets which analysis requires crossing spatial, temporal, thematic and socioeconomic dimensions to enable discoveries of daily urban mobility patterns. This way, information visualization is a suitable approach to support the analysis of urban mobility data, since analysts do not have to learn sophisticated methods to interpret the data visualizations that come to reinforce their cognition and enable the discovery of unstructured insights within the data. 
Thereby, we propose a visualization framework to assist the analysis of urban mobility through indicators describing complementary objects of interest within the data that allow to address three categories of questions underlying the urban mobility phenomenon. A first question seeks to understand the daily traveling routine of a population and the resulting processes of exchange between places, which can be studied through the exploration of amounts, modalities, direction, and variation of travel flows and trips according to different socioeconomic aspects of individuals and land types. A second questioning concerns the temporal variation of population presence throughout a territory, which allows to understand the use of distinct locations by taking into account the socioeconomic characteristics of the people visiting it and the activities they carry out there. The third question seeks to explain the individuals' need of traveling by studying the temporal ordering of trips and activities of individuals (i.e. daily trajectories) within the spatial context of the territory. 
Our framework supports the derivation and visual exploration of indicators describing the territory, travel flows and trips, and daily trajectories, over multiple spatio-temporal resolutions and thematic attributes. Our visualization interface allows to disperse visual representations over multiple analytical displays, enabling users to customize the spatial arrangement of visualizations and indicators in meaningful ways according to the ongoing analysis. Furthermore, we propose a movement-based interaction based on the tilting of a tablet that allows to explore the temporal variation of indicators leveraging tactile and tangible input. The conception of our visualization approach followed an interactive evaluation process that consists of successive user-based evaluations aiming to refine a prototype in order to achieve user performance and satisfaction.