This document discusses opportunities and challenges for data physicalization. It begins by defining data physicalization as physical artifacts whose geometry or properties encode data to support cognition, communication, learning, and decision making. It explores why physicalizing data can engage people by representing data in a familiar, tangible way. Challenges discussed include how to design physical variables, limitations of current 3D printing and shape-changing technologies, and how to support interactivity and animation. The document concludes by calling for connecting communities around data physicalization and establishing resources like an online wiki.

1. Opportunities and Challenges
for Data Physicalization
Yvonne Jansen, Pierre Dragicevic, Petra Isenberg, Jason Alexander,
Abhijit Karnik, Johan Kildal, Sriram Subramanian, Kasper Hornbæk

2. Data Physicalization
… a physical artifact whose geometry or material properties encode data.
… a research area that examines how computer-supported, physical
representations of data (i.e., data physicalizations), can support cognition,
communication, learning, problem solving, and decision making.

3. Hans Rosling, statistician and data communicator

5. Hans Rosling, TED Cannes, 2010

7. Physical Charts for the Tenison Road project, Microsoft Research @ CHI'15

8. Early Data Visualizations
Mesopotamian Clay Tokens (from 5500 BCA)
[Image: Denise Schmandt-Besserat]

9. Early Data Visualizations
Peruvian Quipus (from 2600 BCA)
[Image: Museo Nacional, Lima, Peru]

10. Why physicalize?

11. "Physical props reach the
mind of a large part of the
public because that's how you
represent data in daily life at
the kitchen table, at the cafe
and in the bar."
photobyMartinKjellberg,2013

12. Of all the people in the world (Stan’s Cafe, 2006)
Population Density (Plan B Architecture & Urbanism, 2013)
see dataphys.org/list!
for many more examples
CPI / Cost of Living
(Loren Madsen, 1995)
Data Sculptures

13. Engaging People With Data
Data-Things (Nissen and Bowers, CHI’15)
TastyBeats (Khot et al., CHI’15)
Tension Road Project
(Taylor et al., CHI’15)

14. Why physicalize?
To Engage People With Data

15. 1865 – Hofmann’s Croquet Ball Models
1945 – Electron Density Map and
Molecular Model of Penicillin
Education
Researcher Tools

16. ectron Density Map and
ular Model of Penicillin
1953 – Watson and Crick’s 3D Model of DNA

17. of DNA
1957 – Max Perutz’ Protein Visualizations

18. 2004 – Molecular Graphics Lab at Scripps

19. To Think Through Manipulation
To Engage People With Data
Why physicalize?

20. Intrinsic Support of Active Perception

21. LedermanandKlatzky,1987(link)
Active Perception
Action enriches perception

22. Active Perception
Action enriches perception
Vogt and Magnussen 2007 (link)
Eye movements of a layperson Eye movements of an artist

23. Valdis Krebs (link)
Active Perception
Action enriches perception

24. Photoappaloosa(link)
Active Perception
Action enriches perception
PhotoBlmtduddl(link)

25. Bret Victor (link)
Active Perception
Action enriches perception

26. Active Perception
Walkable Age Pyramid (Atelier Brückner, 2013)
Active Perception

27. Active Perception

28. Why physicalize?
To Think Through Manipulation
To Engage People With Data
To Intrinsically Support Active Perception
more in the paper:
To Address Multiple Senses
To Promote Accessibility
…

29. Challenges

30. How to Design Data Physicalizations?
Challenges

31. Designing physical data representations
position length position + length
examples of visual variables

32. Designing physical data representations
weight
flexibility
temperature
Physical Variables
& many more…
friction

33. Designing physical data representations
Physical Variables
?
Visual Variables

34. Designing physical data representations
How to render physical variables?
3D Printing:
Full color

35. Designing physical data representations
How to render physical variables?
3D Printing:
Different choices of material

36. Designing physical data representations
How to render physical variables?
3D Printing:
Flexible materials

37. Designing physical data representations
How to render physical variables?
3D Printing:
Multi-Material

38. Designing physical data representations
Limitations of current technology
• Limited number of materials
• Cumbersome and slow process
• How to render taste, smell, etc?

39. Designing physical data representations
Limitations of current technology
• Free-floating
objects
Zero-N
(Lee et al, 2011)
Acoustic levitation
(Seah et al, 2014)

40. Supporting Animation and Interactivity

41. Built-in physical interactions
Supporting animation and interactivity
Microsonic Landscapes (Escalante, 2012)
Global Cities (Burdett et al, 2007)

42. Built-in physical interactions
Supporting animation and interactivity
Rearrangeable 3D bar chart (Jansen et al, 2011)

43. Built-in physical interactions
Rearrangeable wooden model of brain scan (Neil Fraser, 2008)
Supporting animation and interactivity

44. Limitations of physical interaction
Supporting animation and interactivity
• Limited functionality
• Can be slow and cumbersome
• Potentially repetitive

45. Synthetic interaction
Supporting animation and interactivity
Sensing Computation Actuation

46. Synthetic interaction
Supporting animation and interactivity
Sensing Computation Actuation

47. Supporting animation and interactivity
Dangling String (Jeremijenko, 1995)
Self-actuated physicalizations

48. Physical Charts, MSR, 2014 The HIVE Big Data Think Tank, Palo Alto, 2013
Supporting animation and interactivity
Self-actuated physicalizations

49. inForm (Follmer, Leithinger et al, 2013) BMW kinetic sculpture (Art+Com, 2008)
Supporting animation and interactivity
Shape displays

50. Shape-change design space (Rasmussen et al, 2012)
Supporting animation and interactivity
Shape displays
Shape displays

51. Limitations of current shape displays
Supporting animation and interactivity
• Limited resolution
• Limited possible forms
• Limited availability
• No actuation of color, texture, taste, etc.
Programmable matter (2060?)

52. Hybrid physical / virtual displays
Supporting animation and interactivity
Emoto, Stefaner et al., 2012
Current Work-Arounds:

53. Hybrid physical / virtual displays
Supporting animation and interactivity
Emoto, Stefaner et al., 2012
Current Work-Arounds:

54. Synthetic interaction
Supporting animation and interactivity
Sensing Computation Actuation

55. Supporting animation and interactivity
virtual
overview
!
physical
detail view
Tangible CityScape (Tang et al., 2013)
Overview and details

56. Supporting animation and interactivity
Changing data representation
Tangible CityScape (Tang et al., 2013)

57. Supporting animation and interactivity
Changing data representation
Tangible CityScape (Tang et al., 2013)

58. Supporting animation and interactivity
EMERGE (Taher et al., CHI’15)
Details on demand

59. Supporting animation and interactivity
Removing data points
EMERGE (Taher et al., CHI’15)

60. Supporting animation and interactivity
Removing data points
EMERGE (Taher et al., CHI’15)

61. Supporting animation and interactivity
physical interactions synthetic interactions
+

62. What to evaluate?
(and how?)

63. Evaluation-Specific Challenges
Assessing the benefits of physicality
• Many modalities to compare with
Paper Virtual reality setupsComputer display

64. Evaluation-Specific Challenges
Going beyond time and error metrics
Walkable Age Pyramid
Atelier Brückner, 2013
• Insights
• Engagement
• Memorability
• Understanding
• Emotional response
• Decision making
• …
Hans Rosling, 2014Rahul Bhargava’s
workshop, 2013

65. Conclusion
Opportunities Challenges
Benefits

66. Conclusion
Opportunities Challenges
Design Implementation EvaluationBenefits

67. http://dataphys.org
Data Physicalization Wiki
• connect the communities
• academia
• artists and designers
• industry
• bibliography
• emerging technologies
http://yvonnejansen.me