Information visualization: presentation
Upcoming SlideShare
Loading in...5

Information visualization: presentation






Total Views
Views on SlideShare
Embed Views



1 Embed 41 41



Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
Post Comment
Edit your comment

Information visualization: presentation Information visualization: presentation Presentation Transcript

  • 13/03/14 pag. 1 Information visualization lecture 4 presentation Katrien Verbert Department of Computer Science Faculty of Science Vrije Universiteit Brussel
  • 13/03/14 pag. 2 Support  for  report  prepara+on.  Many  sources  of  content  are  visible  and  ready  to  hand   A problem
  • 13/03/14 pag. 3 The presentation issue present (tr.v): to offer to view; display.
  • 13/03/14 pag. 4 overview Space  limita+ons     •  Scrolling   •  Overview  +  detail   •  Distor+on   •  Suppression   •  Zoom  and  pan           Time  limita+ons     •  Rapid  serial  visual   presenta+on   •  Eye-­‐gaze                
  • 13/03/14 pag. 5 Space limitations
  • 13/03/14 pag. 6 7.1 A PROBLEM Many of us have found ourselves with a report that has to be completed by a deadline, with the result (Figure 7.1) that the dining room table, extended to its 12- guest state, is covered by piles of paper as well as reports, books, clippings and slides; perhaps with more arranged on the floor and on a couple of chairs. There may even be piles on top of piles. Such a presentation of vital information makes a lot of sense: everything relevant is to hand (hopefully!) and, moreover, its very visibility acts as a reminder (Bolt, 1984, page 2) of what might be relevant at any particular juncture, possibly triggering a situated action (Suchman, 1987). In this environment I can concentrate on creative tasks rather than organisation. Despite the availability of high-resolution displays and powerful workstations I still write most of my reports in this way. Why? Because the display area provided by the typical workstation is far too small to support, visibly, all the sources that are relevant to my composition. 7.2 THE PRESENTATION PROBLEM I am not alone in the sense of having too much data to fit onto a small screen. A very large and expensive screen, for example, would be needed to display the London Underground map in sufficient detail(Figure 1.1), and it would be difficult or impossible to present, on a normal display, the complete organisation chart of IBM or ICI. Moreover, the recent emergence of small and mobile information and communication devices such as PDAs and wearable displays has additionally identified a pressing need for a solution to the ‘ too much data, too little display area’ problem: the presentation problem. How can it be solved, mindful of the need to support the activity of visualising the underlying data? 7.2.1 Scrolling An obvious solution is to scroll the data into and out of the visible area. In other words, to provide a means whereby a long document can be moved past a window until it reaches the required ‘page’ (Figure 7.2). This mechanism is widely used, but carries with it many penalties. One relates to the "Where am I?" problem: I’m working on Chapter 2, (it may be section 2.3, I don’t know) and I want to remind myself of a figure that is in chapter 5, it may be in section 5.3 – or was it 5.6? All I can do is operate the scrolling mechanism and look out for the figure I need, albeit assisted by various cues such as the page number indicated in the scrolling mechanism. With a scrolling mechanism, most of a document is hidden from view. I have the same problem when using a microfilm reader, with the additional complication that if I move the tray to the left, the image moves to the right. A similar difficulty applies to my use of the famous London ‘AtoZ’ street directory. I’m driving along a road that goes off the edge of the page, so I desperately need whatever page contains the continuation of that road (and quickly!). Even if I get it, I will typically have trouble locating the same road on the new page. These and other similar problems can be ameliorated by the provision of context. Much of this chapter, in fact, is concerned with deciding how to provide context. Scrolling
  • 13/03/14 pag. 7 Overview + detail
  • 13/03/14 pag. 8 Source:  Courtesy  Colin  Grimshaw   Overview + detail
  • 13/03/14 pag. 9 Overview + detail hDp://    
  • 13/03/14 pag. 10 A  journey  north  towards  Halifax  requires  detail  of  the  town  (Huddersfield)  through  which   the  traveller  passes   Overview + detail
  • 13/03/14 pag. 11 The  use  of  a  real  or  digitally  simulated  magnifying  glass  masks  detail  around  the  magnified  region     Overview + detail
  • 13/03/14 pag. 12 The  DragMag  technique  allows  flexible  posi+oning  of  the  region  to  be  magnified     Overview + detail
  • 13/03/14 pag. 13 Connection between the detail and overview presentations missing Overview + detail Issues?  
  • 13/03/14 pag. 14 Focus + context
  • 13/03/14 pag. 15 Metaphor  illustra+ng  the   principle  of  the  Bifocal  Display   (a) An information space containing documents, emails, etc. (b) The same space wrapped around two uprights. (c) Appearance of the information space when viewed from an appropriate direction direction of view Distortion
  • 13/03/14 pag. 16 An early illustration of the bifocal display principle
  • 13/03/14 pag. 17 An early illustration of the bifocal display principle
  • 13/03/14 pag. 18 Bifocal display features 1.  Distortion: available display area is allocated to two different regions –  Focus  (undistorted)   –  Context  (distorted)   2.  Information moves smoothly and continuously from context to focus 3.  Display affords for representation –  opportunity  to  use  two  dimensions   –  for  instance,  +me  assigned  to  horizontal  axis   –  type  of  item  to  Y-­‐axis   4.  Main purpose –  Focus:  provide  detail   –  Context:  awareness  and  iden6fica6on     5.  Manual control
  • 13/03/14 pag. 19 What is the Bifocal Display Doing? Transforming the information space to the display space 7.19   Informa+on   space   Display   Space   Normal display Informa+on   space   Display   Space   Bifocal   display   context   focus   Slide  source:  Ken  Brodlie  
  • 13/03/14 pag. 20 A  sequence  of  amino  acids  within  a  protein   Source:  Courtesy  of  Tom  Oldfield   Applications of distortion technique
  • 13/03/14 pag. 21 Table lens without distortion
  • 13/03/14 pag. 22 Table lens with distortion
  • 13/03/14 pag. 23 Table Lens: demo hDp://    
  • 13/03/14 pag. 24 Schematic representation of X-distortion
  • 13/03/14 pag. 25 Schematic representation of combined X- and Y- distortion
  • 13/03/14 pag. 26 hDp://    
  • 13/03/14 pag. 27 Distorted presentation of the London Underground map
  • 13/03/14 pag. 28 11Sun 12 Mon 13 Tue 14 Wed 15 Thur 16 Fri 17Sat Fly LA Kathy to airport Model Maker Check slides, notes. Family barbeque Fly LHR Kathy to collect Chapter 2/ see Dave March JulyJuneMayAprilMar Aug Sept Oct Flight to SFO Tutorial set-up Tutorial United flight Heathrow Pointer Color OHs Jane+John Call Kathy Combined X- and Y-distortion provides a convenient calendar interface
  • 13/03/14 pag. 29 Visual  designer’s  sketch  of  the  applica+on  of  the  flip-­‐zoom  technique  to  the  presenta+on  of   photographs  on  a  Nokia  mobile  phone   Source:  Courtesy  Ron  Bird  
  • 13/03/14 pag. 30 Source:  Courtesy  David  Baar,  IDELIX  SoFware  Inc.   Distorted map on a PDA, showing the continuity of transportation links
  • 13/03/14 pag. 31 Source:  Courtesy  IDELIX  and  Mitsubishi   Distorted map on a table
  • 13/03/14 pag. 32 Equal X- and Y-distortion centred around a manually chosen location in the Macintosh OSX ‘dock’
  • 13/03/14 pag. 33 The Perspective Wall applies a 3D effect to the bifocal display
  • 13/03/14 pag. 34 Advantages Perspective Wall •  User can adjust ratio of detail to context •  Smooth animation helps user perceive object constancy •  Relationship between detail and context is consistent: objects bend around the corner Slide  source:  Ken  Brodlie  
  • 13/03/14 pag. 35 Perspective Wall Perspective gives smoother transition from focus to context Informa+on   space   Display   Space   Perspective Wall context focus Slide  source:  Ken  Brodlie  
  • 13/03/14 pag. 36 overview Space  limita+ons     •  Scrolling   •  Overview  +  detail   •  Distor+on   •  Suppression   •  Zoom  and  pan           Time  limita+ons     •  Rapid  serial  visual   presenta+on   •  Eye-­‐gaze                
  • 13/03/14 pag. 37 Suppression •  Applies a distance function and relevance function •  Less relevant other items are dropped from the display •  Classic example: New Yorker’s idea of the world
  • 13/03/14 pag. 38 Suppression •  Originally proposed by Furnas (1986), but many variations of applications. •  Basic idea: more relevant information presented in great detail; the less relevant information presented as an abstraction. •  Relevance is computed on basis of the importance of information elements and their distance to the focus.
  • 13/03/14 pag. 39 Degree of interest (DOI) function: DOI(a|.=b)  =  API(a)  –  D(A,b)   •  DOI(a|.=b):  DOI  of  a,  given  the  current  focus  is  b.   •  API(a):  sta+c  global  a  priori  importance  measure.   •  D(a,b):  distance  between  a  and  b.  
  • 13/03/14 pag. 40 G P President S M N F K The organization tree of a company
  • 13/03/14 pag. 41 1 2 3 3 4 4 22 1 1 1 22 P Focus Showing the ‘distance’ of each node from the focus of attention
  • 13/03/14 pag. 42 Focus Context P S M NK The context defined by setting an upper threshold of unity for distance from a focus
  • 13/03/14 pag. 43 Example of a display that might be associated with the focus and context
  • 13/03/14 pag. 44 Each  node  in  the  organiza+on  tree  has  been  assigned  an  a  priori  importance  (API)       10 9 9 8 7 7 7 8 8 6 8 8 6 9 API
  • 13/03/14 pag. 45 Degree of Interest (DoI) DoI = API – D Expressed as a function of two quantities: •  A priori importance (API) •  Distance (D) between an item and the item currently in focus
  • 13/03/14 pag. 46 Segng  a  lower  limit  of  6  for  DoI  iden+fies  the  nodes  within  the  shaded  region   8 6 6 8 6 6 6 4 4 4 6 6 4 8 Focus Context Nodal values of degree of interest (=API – D)
  • 13/03/14 pag. 47 Part  of  an   engineering  drawing   Applications of DoI concept
  • 13/03/14 pag. 48 The  engineering  drawing   simplified  in  the  context   of  a  suspected  fault   Applications of DoI concept
  • 13/03/14 pag. 49 Illustra+ng  the  concept  of  a  magic  lens.  (a)  shows  a  conven+onal  map  of  an  area,  (b)  shows  the   loca+on  of  services  (gas,  water  and  electricity  pipes)  in  the  same  area,  and  (c)  a  (movable)  magic   lens  shows  services  in  an  area  of  interest,  in  context   Application in magic lens technique
  • 13/03/14 pag. 50 hDp://    
  • 13/03/14 pag. 51 A  molecular  surface  of  the  protein  transferase  coloured  by  electrosta+c  poten+al  bound  to   DNA  shown  as  a  schema+c.  (ID  =  10mh).  The  magic  lens  window  allows  a  view  of  the   atomic  structure  bonding  to  be  shown,  with  the  bound  ligand  structure  highlighted  as   cylinders,  thereby  providing  a  view  inside  the  protein   Source:  By  kind  permission  of  Tom  Oldfield  and  Michael  Hartshorn   Magic lens
  • 13/03/14 pag. 52 A 3D Flexible and Tangible Magic Lens in Augmented Reality  
  • 13/03/14 pag. 53     A  combina+on  of  rubber-­‐sheet  distor+on  and  suppression  lead  to  a  map  appropriate  to  a   journey  from  one  city  to  another   Combined distortion and suppression
  • 13/03/14 pag. 54 The rubber-sheet distortion technique employed in the map
  • 13/03/14 pag. 55 Historical note •  Distortion and suppression appeared in early 1980s •  Need to maintain a balanced view of focus + context identified earlier – for example by Farrand (1973) “an effective transformation must somehow maintain global awareness while providing detail” “… there is a need for presenting a display with 1. sufficient detail for interaction, while 2. maintaining global vision of the entire scene.”
  • 13/03/14 pag. 56 Fisheye view •  Farrand also coined the term “fisheye” •  Nowadays appears to refer to both distortion and suppression
  • 13/03/14 pag. 57 Fisheye Menus •  Here is the same idea applied to menus –  Ben  Bederson,  University  of  Maryland   •  See also: –  hDp://­‐demo.shtml   ENV  2006  
  • 13/03/14 pag. 58 Fisheye View, Polyfocal Display Can  distort  boundaries  because  applied  radially  rather  than  x  y   1D  Fisheye   2D  Polyfocal   Slide  source:  Hornung  and  Zagreus    
  • 13/03/14 pag. 59 hDp:// cmsc838s/+chi/fisheye.html    
  • 13/03/14 pag. 60 hDp://  
  • hDps://    
  • 13/03/14 pag. 62 Source:  By  kind  permission  of  Patrick  Baudisch   The use of representation (by a ‘halo’) to provide context for a small display
  • 13/03/14 pag. 63 overview Space  limita+ons     •  Scrolling   •  Overview  +  detail   •  Distor+on   •  Suppression   •  Zoom  and  pan           Time  limita+ons     •  Rapid  serial  visual   presenta+on   •  Eye-­‐gaze                
  • 13/03/14 pag. 64 Panning  is  the  smooth  movement  of  a  viewing  frame  over  a  2D  image     Zoom and pan
  • 13/03/14 pag. 65   Zooming  is  the  increasing  magnifica+on  of  a  decreasing  frac+on  of  an  image  (or  vice  versa)   Zoom and pan
  • 13/03/14 pag. 66 Zooming •  Conventional zooming-in –  No  change  in  data  or  representa+on  –  only  filtering   –  Loss  of  context     •  ≠distortion whose purpose is to permit focusing rather than filtering •  Supports two cognitive tasks (Cairns and Craft 2005) –  Zooming-­‐in:  extraneous  informa+on  is  removed  from  visual  field  –  more   manageable  view   –  Zooming-­‐out:  reveals  hidden  informa+on  
  • 13/03/14 pag. 67 A space-scale diagram relevant to combined zooming and panning Furnas  and  Bederson  (1995)  
  • 13/03/14 pag. 68 Google earth
  • 13/03/14 pag. 69 Exploring Information Spaces by Using Tangible Magic Lenses hDp://­‐mF4_OAhU0    
  • 13/03/14 pag. 70 Semantic zoom •  Previous example: geometric zoom –  Con+nuous   –  Zooming-­‐in:  filtering  and  loss  of  context   •  Semantic zoom –  Discrete  transi+on   –  Addi+onal  detail  
  • 13/03/14 pag. 71 A combination of geometric and semantic zoom
  • 13/03/14 pag. 72 overview Space  limita+ons     •  Scrolling   •  Overview  +  detail   •  Distor+on   •  Suppression   •  Zoom  and  pan           Time  limita+ons     •  Rapid  serial  visual   presenta+on   •  Eye-­‐gaze                
  • 13/03/14 pag. 73 A  collec+on  of  images  is  presented,  one  at  a  +me,  at  a  rapid  rate  (e.g.,  ten  per  second)     time Rapid serial visual presentation
  • 13/03/14 pag. 74 Tile mode: concurrent presentation of images opposed to ‘slide show mode’
  • 13/03/14 pag. 75 ‘Floa+ng  RSVP’  in  which   images  appear  to  approach   the  viewer  from  a  distance.   Sensi+ve  arrows  allow  the   speed  and  direc+on  of   ‘movement’  to  be  controlled   by  a  user   Source:  Courtesy  Kent  WiNenburg   Floating RSVP
  • 13/03/14 pag. 76 The  contents  of  an  online  bookstore  are  presented  in  ‘collage  mode’  RSVP,  simula+ng  the   placing  of  book  covers  on  a  table  in  sequence.  The  set  of  arrows  just  under  the  presenta+on   allows  control  of  the  speed  and  direc+on  of  presenta+on   Source:  Courtesy  Kent  WiNenburg   Collage mode RSVP
  • 13/03/14 pag. 77 An  interface  facilita+ng  the  browsing  of  posters  adver+sing  videos.    Cursor  movement  along  the   stacks  causes  posters  to  briefly  ‘pop  out’  sideways,  and  the  whole  bifocal  structure  can  be   scrolled  to  bring  a  video  of  interest  to  the  central  region,  where  a  mouse  click  will  cause  a  clip   from  a  video  to  be  played  (Lam  and  Pence  1997)   RSVP + bifocal principle
  • 13/03/14 pag. 78 Space-time trade-off
  • 13/03/14 pag. 79 Space-time trade-off
  • 13/03/14 pag. 80   An  experiment  to  test  a  subject’s  ability  to  recognise  the  presence  or  absence  of  a  previously   viewed  target  image  within  a  collec+on  presented  sequen+ally  at  a  rate  of  around  ten  per  sec.     Prior instruction to subject Subjectsʼ performance “Here is a target image. Tell me if this image appears in the sequence of N images youʼre about to see” Recognition about 80% to 90% successful time about 100 ms unrelated images Presentation of images Briefly glimpsed images
  • 13/03/14 pag. 81 Representa+on  of  limits  on  display  area  and  total  presenta+on  +me  by  a  ‘resource  box’     Display area Presentation time Space and time resources
  • 13/03/14 pag. 82 Source:  Courtesy  of  Katy  Cooper   Three ‘static’ image presentation modes (A, B, C) and three ‘moving’ image presentation modes (D, E, F)
  • 13/03/14 pag. 83 Slide show
  • 13/03/14 pag. 84 Mixed
  • 13/03/14 pag. 85 Tile
  • 13/03/14 pag. 86 Diagonal
  • 13/03/14 pag. 87 Ring
  • 13/03/14 pag. 88 Stream
  • 13/03/14 pag. 89 Source:  Courtesy  of  Katy  Cooper   Favorite mode?
  • 13/03/14 pag. 90 The  accuracy  with  which  the  presence  or  absence  of  a  target  image  was  reported  for  the   six  presenta+on  modes,  averaged  over  all  tasks  and  presenta+on  +mes.         1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Slide-show Mixed Tile Diagonal Ring Stream Recognition accuracy Presentation modes
  • 13/03/14 pag. 91 The  (sta+c)  slide-­‐show,  mixed  and  +le  image  presenta+on  modes  account  for  three-­‐quarters  of   the  preferred  modes  (Cooper  et  al.  2006)  
  • 13/03/14 pag. 92 Almost  all  the  least  preferred  image  presenta+on  modes  were  moving  modes  and  the  stream   mode  accounted  for  over  half    
  • 13/03/14 pag. 93   A  simple  representa+on  of  eye-­‐gaze  behaviour.  The  rapid  saccades  are  shown  green,  the   fixa+ons  (F)  of  varying  dura+on  by  circles  of  propor+onate  size   F F F F F F F F Eye-gaze
  • 13/03/14 pag. 94 Eye-gaze trajectory slide show
  • 13/03/14 pag. 95 Eye-gaze trajectory tile mode
  • 13/03/14 pag. 96 Eye-gaze trajectories mixed mode
  • 13/03/14 pag. 97 Eye-gaze trajectories diagonal mode for subject who disliked the mode
  • 13/03/14 pag. 98 Eye-gaze trajectories diagonal mode for subject who liked the mode
  • 13/03/14 pag. 99 Eye-gaze trajectory stream for subject who disliked the mode
  • 13/03/14 pag. 100 The  acquisi+on  of  an  expanding  target.  (a)  The  dormant  appearance  of  the  image  collec+on,   and  (b)  its  appearance  when  the  cursor  rests  over  image  6     1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 10 11 12 13 14 15 16 17 18 19 209861 2 3 4 (a) (b) Manual control: ‘expanding target’ presentation mode
  • 13/03/14 pag. 101 An  experiment  in  which  a  subject  first  views  the  rapid  (e.g.,  10  per  second)  presenta+on  of  a   collec+on  of  images,  is  then  shown  a  single  image  and  asked  to  say  whether  that  image  was   part  of  the  collec+on.  Iden+fica+on  success  is  highly  dependent  upon  the  +me  elapsing   between  the  end  of  the  presenta+on  and  the  ques+oning  of  the  subject   Prior instruction to subject Presentation of image collection Subject’s performance about 100ms unrelated images time None The subject was shown an image and then asked, ‘Was this image present in the sequence you have just seen?’ Recognition success was 10% to 20% unless the question was aksed within about 4 seconds of the end of the presentation Models of human visual performance
  • 13/03/14 pag. 102 An  experiment  in  which  a  collec+on  of  images  is  presented  to  a  subject.  Each  image  is  presented   briefly  (e.g.,  for  100ms)  and  followed  by  a  ‘visual  mask’  las+ng  about  300ms.  Subjects  were  able   to  say,  with  a  considerable  degree  of  success,  whether  an  image  shown  arerwards  had  been   part  of  the  presenta+on   Prior  instruc+on     to  subject   Presenta+on  of  image  collec+on   Subject’s  performance   about  300ms   unrelated  images   +me   None   The  subject  was  shown  an  image  and  then     asked,  ‘Was  this  image  present  in  the     sequence  you  have  just  seen?’   Up  to  92%  recogni+on  success   etc  .  .  .  .  Visual   mask    .     Visual   mask       Visual   mask   about  100ms   Models of human visual performance
  • 13/03/14 pag. 103 A  third  palleDe  for  the  interac+on  designer,  addressing  issues  of  presenta+on   Presentation   concepts and   techniques   Scrolling   Overview+detail   Distortion   Suppression   Zoom   Pan   RSVP   Eye gaze   Recap
  • 13/03/14 pag. 104 Visual Information Seeking: Mantra Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Ben  Shneiderman,  1996  
  • 13/03/14 pag. 105 Shneiderman’s “7 Tasks” •  Overview task –  overview of entire collection •  Zoom task –  zoom in on items of interest •  Filter task –  filter out uninteresting items •  Details-on-demand task –  select an item or group to get details •  Relate  task   –  relate  items  or  groups  within  the   collec+on   •  History  task     –  keep  a  history  of  ac+ons  to  support   undo,  replay,  and  progressive   refinement   •  Extract  task   –  allow  extrac+on  of  sub-­‐collec+ons  and   of  the  query  parameters  
  • 13/03/14 pag. 106 Questions?
  • 13/03/14 pag. 107 Readings •  Chapter 4
  • 13/03/14 pag. 108 References •  Furnas, G. W., & Bederson, B. B. (1995, May). Space-scale diagrams: Understanding multiscale interfaces. In Proceedings of the SIGCHI conference on Human factors in computing systems (pp. 234-241). ACM Press/Addison- Wesley Publishing Co.. •  Shneiderman, B. (1996, September). The eyes have it: A task by data type taxonomy for information visualizations. In Visual Languages, 1996. Proceedings., IEEE Symposium on (pp. 336-343). IEEE. •  Some relevant notes: presentation.html
  • 13/03/14 pag. 109 Research presentations
  • 13/03/14 pag. 110 Research presentations •  Schedule on PointCarré •  Select a second paper in the same slot for questions: e.g. session 1: •  Links to doodle polls for all six sessions will be included in the schedule.
  • 13/03/14 pag. 111 Team project
  • 13/03/14 pag. 112 Team project milestones 1.  Form teams 2.  Project proposal 3.  Intermediate presentation 4.  Final presentation 5.  Short report due  27  Feb.   due  13  March   due  3  April   22  May   due  29  May  
  • 13/03/14 pag. 113 Project proposal 1 page description of your intended project: –  mo+va+on   –  which  datasets  you  will  use   –  current  status.  If  available,  first  designs.   –  problems/ques+ons   due 13 March If you want earlier feedback, send us your proposal earlier ;-)
  • 13/03/14 pag. 114 Data collection • 1gHwVWHZLzWdSz1F37jA1Gungrl56bT215M6FYW3YqGY/ viewform Or • Anonymous! Choose your own ID. •  Please report your data ;-)