Ph.D
Group : Human-Centered Computing
Embodied Interaction for Data Manipulation Tasks on Wall-sized Displays
Starts on 01/10/2012
Advisor : BEAUDOUIN-LAFON, Michel
[Olivier CHAPUIS]
Funding : Contrat doctoral organisme (EPST, EPA ayant une mission d'enseignement supérieur)
Affiliation : Université Paris-Saclay
Laboratory : LRI-INSITU
Defended on 17/12/2015, committee :
Directeur de thèse :
M. Michel BEAUDOUIN-LAFON, Professeur, Univ Paris Sud
Co-directeurs de thèse :
M. Olivier CHAPUIS, Chargé de Recherche, CNRS
M. Eric LECOLINET, Maître de Conférences, Telecom ParisTech
Rapporteurs:
M. Albrecht SCHMIDT, Professor, University of Stuttgart
M. Laurent GRISONI, Professeur, University of Lille 1
Examinateurs :
Mme Joanna MCGRENERE, Professor, University of British Columbia
M. Jean-Claude MARTIN, Professeur, Univ Paris Sud
M. Nicolai MARQUARDT, Assistant Professor, University College London
Research activities :
Abstract :
Large data sets are increasingly used in various professional domains. This raises challenges in managing and using them for sense-making, searching and classification tasks. Not only does big data require advanced algorithms to process the data, it also needs users' judgment to correct and interpret it. This dissertation explores the use of large, high-resolution wall-sized displays, which can display large amounts of information, to support user interaction with large data sets. It contributes with novel insights on the interactive phenomena found in this context through laboratory experiments, as well as with the design and prototyping of novel interaction techniques for supporting collaboration.
I begin with discussing the user needs for data manipulation with large data sets, as uncovered from interviews with and observations of real users. Then I introduce a series of controlled experiments that study user interaction with large wall-sized displays, in both single-user and collaborative situations. The first experiment shows that physical navigation in front of a large display outperforms virtual navigation on a desktop monitor, because large displays leverage users' whole-body skills to navigate and manipulate data. Another experiment shows benefits of an interaction technique that combines multiple users' actions to issue a command, and provides insights on different collaboration styles.
Based on the empirical insights, I then demonstrate my design of new interaction techniques to explore embodied interaction that leverages users' the physical and social skills in a co-located environment to carry out tasks with the "computer". Two prototypes are built to facilitate data manipulation and exchange between co-workers in various collaborative situations. I show the techniques and findings from preliminary evaluation, discuss the perspectives and directions for future work.
