version=pmwiki-2.2.40 ordered=1 urlencoded=1 agent=Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/29.0.1547.65 Safari/537.36 author= charset=UTF-8 csum= ctime=1378842243 host=129.175.15.11 name=Research.Dynaphysvis2 rev=1 targets= text=

%25center%25 Investigating Dynamic Physical Visualizations

[[

]]

%25center%25 %25height=200px%25

%25center%25 Dynamic art installation consisting in 714 metal spheres hanging from thin steel wires attached to individually-controlled stepper motors (link).

[[

]]

Update: a PhD thesis grant on this topic has been awarded to Mathieu Le Goc. The position is now closed.

Abstract

When designing interactive information visualization systems, we used to have in mind a single user sitting in front of a desktop computer with a mouse and keyboard. With the advent of new input and output technologies, we have to rethink these stereotypes. New interaction paradigms need to be explored that better exploit humans' natural abilities to perceive and manipulate physical objects. Recent research has shown that moving parts of information visualization systems into the physical world can yield many benefits. But this research has so far been focusing on passive physical objects that do not update themselves with new data and do not support elaborate forms of data exploration (e.g., filtering, search, etc.). The goal of this thesis will be to design, build and study dynamic physical visualizations, i.e., physical objects whose shape reflect dynamic data and that can be interactively manipulated to explore large datasets.

Context

Information visualization is a growing topic of research. A study estimated that the amount of data produced in the world was increasing by 50%25 each year [1]. Since our brains and sensory capacities have not changed in the meantime, gaining competitive advantage from all this data depends heavily on the effectiveness with which we support human abilities to perceive, understand, and alter it. The Aviz INRIA research team (www.aviz.fr) seeks to improve analysis and exploration of large and complex datasets with new forms of interactive visualization.

Tangible user interfaces make it possible to interact with information in radically new ways. They have been studied in human-computer interaction but have received little attention in information visualization. In tangible user interfaces, users interact with digital information through physical objects. An early example is the Marble Answering Machine by Durrell Bishop [2,3], where messages are represented by marbles and can be played back by dropping a marble into a dish. Many other systems have been proposed since then, and this research field is still very active [4].

At Aviz, we started to investigate how tangible interaction can facilitate data exploration. We showed how moving visualization controls to the physical world can facilitate data exploration by allowing for locomotion and eyes-free interaction in large display environments [5]. We also showed how moving the visualization itself to the physical world can facilitate information retrieval, especially for 3D visualizations and tasks that are demanding to short-term memory [6].

These initial studies have opened up a new domain at the intersection of information visualization and tangible user interfaces, revealing new opportunities but also new challenges. One major challenge is the design and the study of dynamic -- possibly interactive -- physical visualizations, i.e., physical visualizations whose shape can be dynamically altered to support dynamic datasets, the dynamic tuning of visualizations and interactive exploration. There has been virtually no research on this difficult and yet rich and promising topic.

Objectives

The purpose of this PhD thesis topic is the engineering, design and study of dynamic physical visualizations. Four distinct but complementary directions of research are:

  • Designing and building functional dynamic physical visualizations. Although a few dynamic "data sculptures" have been built by designers and artists as part of museum installations, these are all data-specific and difficult to recognize and read [7]. Conversely, there has been attempts at building dynamic physical counterparts of classic visualizations like 2D bar charts, but they are very limited in data size and in functionality [6,7]. Building fully-functional dynamic physical visualizations pose non-trivial problems of hardware, electronic and mechanical design.
  • Designing new interaction techniques for dynamic physical visualizations. The few existing prototypes of dynamic physical visualizations are either non-interactive or support stereotyped interaction styles that only make a poor use of the human capacity to manipulate physical objects [9]. Interaction with physical visualizations has to be entirely rethought, as interaction paradigms from traditional computer visualizations are likely to be suboptimal. A few concept videos have been proposed that illustrate how humans may interact with programmable matter in the future [10,11] but they focus on CAD tasks, not data exploration tasks. They may however provide an initial source of inspiration.
  • Evaluating dynamic physical visualizations. Previous work has suggested that physical visualizations can have a number of advantages over on-screen visualizations [6,8], but no empirical study exists that involves dynamic physical visualizations. Yet it is very important for the advancement of scientific knowledge to understand what in physical objects makes them more suited to some data exploration tasks and less suited to other types of tasks. Despite almost 20 years of research in tangible user interfaces, the concepts of "physicality" and "tangibility" are still very poorly understood, and little empirical work exists that explores the vast continuum between virtual and physical setups.
  • Building theories and models for dynamic physical visualizations. Information visualization and human-computer interaction lack theories and models, which are also important for the advancement of scientific knowledge. A good conceptual model should be able to capture existing designs (descriptive power), help criticize and compare designs (evaluative power), and help build new designs (generative power) [12]. A useful direction of research would be to develop theories and models that help understand dynamic physical visualizations and where they fit in the design space of interactive information visualizations [9].

Work Program

Given this topic's strong focus on hardware design, ideally the PhD student should initially have a strong background in electronic and hardware design. At the beginning of the thesis, the student will first get acquainted with the fields of information visualization, tangible user interface and user interface evaluation methods through a combination of literature review and by attending modules from our HCI Master class (introduction to HCI + information visualization). Then the student may pick a small research problem that seems novel, interesting and feasible in order to ensure a first publication and acquire the necessary skills for publishing scientific results. A further step will then consist in empirically validating and generalizing the approach. The outcome of the PhD thesis may be a general framework for the design, construction and evaluation of dynamic physical visualizations and/or related artefacts.

Requirements

The candidate must:

  1. Hold a Masters degree in computer science or equivalent
  2. Have sufficient training or experience in:
    • Electronics and hardware engineering
    • Software development
  3. Have a genuine interest in -- and if possible some basic knowledge of:
    • Information visualization
    • Human-computer interaction
    • Digital fabrication
    • Empirical research methods
  4. Be creative
  5. Possess good analytical thinking stills
  6. Have a decent level of oral and written English. Speaking French not required upon arrival.

Advising

The student will be advised by Pierre Dragicevic and Jean-Daniel Fekete. Possible collaboration with Hiroshi Ishii from the Tangible Media Group at MIT.

Links

Additional links

(Added later on)

time=1378842243 title=Investigating Dynamic Physical Visualization author:1378842243= diff:1378842243:1378842243:=1,85d0 

 %25center%25  Investigating Dynamic Physical Visualizations



 [[

]] 

 %25center%25 %25height=200px%25 



 %25center%25  Dynamic art installation consisting in 714 metal spheres hanging from thin steel wires attached to individually-controlled stepper motors (link). 



 [[

]] 

 Update: a PhD thesis grant on this topic has been awarded to Mathieu Le Goc. The position is now closed.



 !Abstract



 When designing interactive information visualization systems, we used to have in mind a single user sitting in front of a desktop computer with a mouse and keyboard. With the advent of new input and output technologies, we have to rethink these stereotypes. New interaction paradigms need to be explored that better exploit humans' natural abilities to perceive and manipulate physical objects. Recent research has shown that moving parts of information visualization systems into the physical world can yield many benefits. But this research has so far been focusing on passive physical objects that do not update themselves with new data and do not support elaborate forms of data exploration (e.g., filtering, search, etc.). The goal of this thesis will be to design, build and study dynamic physical visualizations, i.e., physical objects whose shape reflect dynamic data and that can be interactively manipulated to explore large datasets.



 !Context



 Information visualization is a growing topic of research. A study estimated that the amount of data produced in the world was increasing by 50%25 each year [1]. Since our brains and sensory capacities have not changed in the meantime, gaining competitive advantage from all this data depends heavily on the effectiveness with which we support human abilities to perceive, understand, and alter it. The Aviz INRIA research team (www.aviz.fr) seeks to improve analysis and exploration of large and complex datasets with new forms of interactive visualization.



 Tangible user interfaces make it possible to interact with information in radically new ways. They have been studied in human-computer interaction but have received little attention in information visualization. In tangible user interfaces, users interact with digital information through physical objects. An early example is the Marble Answering Machine by Durrell Bishop [2,3], where messages are represented by marbles and can be played back by dropping a marble into a dish. Many other systems have been proposed since then, and this research field is still very active [4].



 At Aviz, we started to investigate how tangible interaction can facilitate data exploration. We showed how moving visualization controls to the physical world can facilitate data exploration by allowing for locomotion and eyes-free interaction in large display environments [5]. We also showed how moving the visualization itself to the physical world can facilitate information retrieval, especially for 3D visualizations and tasks that are demanding to short-term memory [6].



 These initial studies have opened up a new domain at the intersection of information visualization and tangible user interfaces, revealing new opportunities but also new challenges. One major challenge is the design and the study of dynamic -- possibly interactive -- physical visualizations, i.e., physical visualizations whose shape can be dynamically altered to support dynamic datasets, the dynamic tuning of visualizations and interactive exploration. There has been virtually no research on this difficult and yet rich and promising topic.



 !Objectives



 The purpose of this PhD thesis topic is the engineering, design and study of dynamic physical visualizations. Four distinct but complementary directions of research are:



 * Designing and building functional dynamic physical visualizations. Although a few dynamic "data sculptures" have been built by designers and artists as part of museum installations, these are all data-specific and difficult to recognize and read [7]. Conversely, there has been attempts at building dynamic physical counterparts of classic visualizations like 2D bar charts, but they are very limited in data size and in functionality [6,7]. Building fully-functional dynamic physical visualizations pose non-trivial problems of hardware, electronic and mechanical design.



 * Designing new interaction techniques for dynamic physical visualizations. The few existing prototypes of dynamic physical visualizations are either non-interactive or support stereotyped interaction styles that only make a poor use of the human capacity to manipulate physical objects [9]. Interaction with physical visualizations has to be entirely rethought, as interaction paradigms from traditional computer visualizations are likely to be suboptimal. A few concept videos have been proposed that illustrate how humans may interact with programmable matter in the future [10,11] but they focus on CAD tasks, not data exploration tasks. They may however provide an initial source of inspiration.



 * Evaluating dynamic physical visualizations. Previous work has suggested that physical visualizations can have a number of advantages over on-screen visualizations [6,8], but no empirical study exists that involves dynamic physical visualizations. Yet it is very important for the advancement of scientific knowledge to understand what in physical objects makes them more suited to some data exploration tasks and less suited to other types of tasks. Despite almost 20 years of research in tangible user interfaces, the concepts of "physicality" and "tangibility" are still very poorly understood, and little empirical work exists that explores the vast continuum between virtual and physical setups.



 * Building theories and models for dynamic physical visualizations. Information visualization and human-computer interaction lack theories and models, which are also important for the advancement of scientific knowledge. A good conceptual model should be able to capture existing designs (descriptive power), help criticize and compare designs (evaluative power), and help build new designs (generative power) [12]. A useful direction of research would be to develop theories and models that help understand dynamic physical visualizations and where they fit in the design space of interactive information visualizations [9].



 !Work Program



 Given this topic's strong focus on hardware design, ideally the PhD student should initially have a strong background in electronic and hardware design. At the beginning of the thesis, the student will first get acquainted with the fields of information visualization, tangible user interface and user interface evaluation methods through a combination of literature review and by attending modules from our HCI Master class (introduction to HCI + information visualization). Then the student may pick a small research problem that seems novel, interesting and feasible in order to ensure a first publication and acquire the necessary skills for publishing scientific results. A further step will then consist in empirically validating and generalizing the approach. The outcome of the PhD thesis may be a general framework for the design, construction and evaluation of dynamic physical visualizations and/or related artefacts.



 !Requirements



 The candidate must:

 # Hold a Masters degree in computer science or equivalent

 # Have sufficient training or experience in:

 ** Electronics and hardware engineering

 ** Software development

 # Have a genuine interest in -- and if possible some basic knowledge of:

 ** Information visualization

 ** Human-computer interaction

 ** Digital fabrication

 ** Empirical research methods

 # Be creative

 # Possess good analytical thinking stills

 # Have a decent level of oral and written English. Speaking French not required upon arrival.



 !Advising

 The student will be advised by Pierre Dragicevic and Jean-Daniel Fekete. Possible collaboration with Hiroshi Ishii from the Tangible Media Group at MIT.



 !Links

 * [1] Lyman and Varian (2003) How Much Information 2003?

 * [2] Poynor (1995) The Hand that Rocks the Cradle

 * [3] Yi-Luen Do (2011) Durrell Bishop’s Marble Answering Machine (with video)

 * [4] Ishii et al (2012) Radical Atoms: Beyond Tangible Bits,Toward Transformable Materials.

 * [5] Jansen, Dragicevic and Fekete (2012) Tangible Remote Controllers for Wall-Sized Displays.

 * [6] Jansen, Dragicevic and Fekete (2013) Evaluating the Efficiency of Physical Visualizations.

 * [7] Jansen and Dragicevic (2013) List of Active Physical Visualizations.

 * [8] Vande Moere (2008). Beyond the tyranny of the pixel: Exploring the physicality of information visualization

 * [9] Jansen and Dragicevic (2014) An Interaction Model for Visualizations Beyond The Desktop. Conditionally accepted to infoVis'2014 -- private copy can be sent upon request.

 * [10] Intel Research (2006) Car Design via Claytronics Technology (video)

 * [11] Tangible Media group (2013) Perfect Red (video)

 * [12] Beaudouin-Lafon (2000) Instrumental Interaction: An Interaction Model for Designing Post-WIMP Interfaces



 !Additional links

 (Added later on)



 * The GHOST Fet project (possible interesting collaborations).

 * Rasmussen et al (2012) Shape-Changing Interfaces: A Review of the Design Space and Open Research Questions

 * Actuation and Tangible User Interfaces: the Vaucanson Duck, Robots, and Shape Displays

 * DESIGNING KINETIC INTERACTIONS FOR ORGANIC USER INTERFACES

 * Shape-changing matter terminology (Google Doc)

\ No newline at end of file

host:1378842243=129.175.15.11