newm n328 visualizing information
play

NEWM - N328 Visualizing Information Khairi Reda | redak@iu.edu - PowerPoint PPT Presentation

NEWM - N328 Visualizing Information Khairi Reda | redak@iu.edu School of Informa5cs & Compu5ng, IUPUI Welcome! This course is about Data Visualiza1on Defining Visualization The transforma5on of data into visual representa1ons to aid


  1. NEWM - N328 Visualizing Information Khairi Reda | redak@iu.edu School of Informa5cs & Compu5ng, IUPUI

  2. Welcome! This course is about Data Visualiza1on

  3. Defining Visualization • The transforma5on of data into visual representa1ons to aid people in the analysis, explora5on, and communica5on of that data

  4. - data - visual representa5ons - people - computers

  5. - data - visual representa5ons - people - computers

  6. Healthcare

  7. Financial markets

  8. Scien1fic instruments: Hubble Space Telescope

  9. Social Media

  10. How much data are we generating?

  11. In 2010: 1,200 exabytes 1 exabyte = 1,000,000,000,000,000,000 bytes 1 exabyte = 10 18 bytes 10x increase every 5 years Based on slide by Jeff Heer

  12. The problem Web, Books, Papers, Scientific data, How? News, Product reviews, Sight, … Data Hearing, Touch, Smell, Taste, Telepathy? Based on a slide by John Stasko / Chris North

  13. Why use vision to analyze information? 1 . Vision is the highest bandwidth channel into our brain

  14. Sensory bandwidth T. Norretranders , The User Illusion: Cutting Consciousness Down to Size, 1999 How much we are ac5vely aware of

  15. Why use vision to analyze information? 1 . Vision is the highest bandwidth channel into our brain 2 . Cogni5on is limited. Visual percep5on beats cogni5on

  16. 345OIJDFG98C90U5ET09VBKK23490XIVBCIBJ0345T09U 2G84GDF09U34590IDFK90345I-09345K90FU90DF90JDF 34T09X90DFJG90J34T09J34509J3459DFG08JKLSTJP435 DFDFG45OJERPOTJ45OPIJFDGLKM34T5XJSCTYY7K456 POJ345OIJLGJKOPE390UVFHUDGH9345H9R4N97HWTIO MADSIOPEJDFGPJ4309UT509345PODFGX093490823JFD PWDEIJ3408UDFMV984385Y0834N92384YU8DFB0H3T4N 345J09JDFG09J345X98U5Y09JGFB089H34509UJ45TM0IG P5JDGIOEGWJPIO345U345OPIJDTOPI3458345JPODFG09 45POJ34X09345J08EFJ825HJDFSJIPADOPQWIXERWNVF Based on a slide by John Stasko

  17. 345OIJDFG98C90U5ET09VBKK23490 X IVBCIBJ0345T09U 2G84GDF09U34590IDFK90345I-09345K90FU90DF90JDF 34T09 X 90DFJG90J34T09J34509J3459DFG08JKLSTJP435 DFDFG45OJERPOTJ45OPIJFDGLKM34T5 X JSCTYY7K456 POJ345OIJLGJKOPE390UVFHUDGH9345H9R4N97HWTIO MADSIOPEJDFGPJ4309UT509345PODFG X 093490823JFD PWDEIJ3408UDFMV984385Y0834N92384YU8DFB0H3T4N 345J09JDFG09J345 X 98U5Y09JGFB089H34509UJ45TM0IG P5JDGIOEGWJPIO345U345OPIJDTOPI3458345JPODFG09 45POJ34 X 09345J08EFJ825HJDFSJIPADOPQWI X ERWNVF Based on a slide by John Stasko

  18. Given these 50 numbers what number appears most oQen? Slide by Miriah Meyer, University of Utah http://www.cs.utah.edu/~miriah/teaching/cs6630/

  19. Given these 50 numbers what number appears most oQen? Slide by Miriah Meyer, University of Utah http://www.cs.utah.edu/~miriah/teaching/cs6630/

  20. How has the unemployment rate in the US changed since 2006? Seasonally adjusted unemployment rate in the US Bureau of Labor Statistics

  21. How is the unemployment rate in the US changing since 2006?

  22. Why use vision to analyze information? 1 . Vision is the highest bandwidth channel into our brain 2 . Cogni5on is limited. Percep5on beats cogni5on 3 . Working memory is limited; external representa5ons expand our memory

  23. 4 26 57 x 26 x 57 2 18 130 + 1482

  24. The world is its own memory “It’s things that make us smart” -Don Norman http://www.femside.com/wp-content/uploads/2013/12/desk-notes-work-office-computer.jpg

  25. Why use vision to analyze information? 1 . Vision is the highest bandwidth channel into our brain 2 . Cogni5on is limited. Percep5on beats cogni5on 3 . Working memory is limited; external representa5ons expand our memory

  26. Why use vision to analyze information? 1 . Vision is the highest bandwidth channel into our brain 2 . Cogni5on is limited. Percep5on beats cogni5on 3 . Working memory is limited; external representa5ons expand our memory 4 . Visuals are an integral part of our culture

  27. • “I see what you’re saying” • “Seeing is believing” • “I now see the big picture” • “A picture is worth a thousand words” Based on a slide by John Stasko

  28. - data - visual representa1ons - people - computers

  29. - data - visual representa5ons - people - computers

  30. Why involve people in the analysis of data? • Computers are very good at compu5ng an answer to a specific ques1on • They are less useful good when we do not know what we are looking for in advance • How does the employment market react to changes in interest rate? • What is the effect of gene muta5on on cancer risk? • Computers are bad at “hunches”

  31. - data - visual representa5ons - people - computers

  32. - data - visual representa5ons - people - computers

  33. Why involve computers in the analysis of data? • Efficiency: Process large quan55es of informa5on quickly • Interact with the data; zoom, filter, switch views, details on demand • Quality: precise representa5on of the data • Re-use of code for different datasets

  34. - data - visual representa5ons - people - computers

  35. A little bit of history…

  36. Bailee Clift

  37. E. W. Gilbert . Simplified version of Snow’s map: http://www.ph.ucla.edu/epi/snow/cartographica39%284%291_14_2004.pdf

  38. Charles Menard. Napoleon’s 1812 campaign to Moscow

  39. How to use visualization Communication Analysis •When we know the facts •When we don’t know and want to show them what to look for to other people •When we don’t have a •When we want to tell a priori question story with data

  40. What we will learn in this course • Understand when, why, and how to use visualiza5on for the analysis of data • We will learn about a variety of visualiza5ons for a range of data types • You will gain experience with modern visualiza5on tools to create your own visualiza5ons

  41. Visual Percep1on and Cogni1on Collin Ware

  42. Visual Marks & Channels Semiology of Graphics J. Bertin

  43. InnoVis group at U Calgary Visual representa1ons Mike Bostock

  44. Maps and Geo-visualiza1ons Mike Bostock

  45. Mike Bostock Networks

  46. UC Berkeley Visualization Lab Yiwen Sun Interac1on techniques

  47. Required reading Interac1ve Data Visualiza1on Analysis Visualiza1on for the and Design Web Tamara Munzner Scoc Murray

  48. http://khreda.com/teaching/2020/N328

  49. Grading Percentage Assignment 1 Intro to Tableu 15% Assignment 2 Exploratory Visual Analysis 20% Visualization Rolling basis — starting from week 4 25% Critique Final Project (group) 30% Participation in paper discussion and project critiques 10%

  50. Visualization Critique • Every week (star5ng from week 4), 2-3 students will be responsible for selec5ng a visualiza5on and pos5ng a design cri5que on Canvas • You are responsible to post one design cri5que during the semester: • Find a visualiza5on from news web sources, textbooks, or scien5fic literature • Explain the data being shown • Describe the visual encoding used in the visualiza5on: put into words what the visualiza5on is trying to show and how. • Cri5que the visualiza5on: what works, and what doesn’t? Is the visualiza5on clear, or is it misleading? • Do you like it? Why? What would you do to improve it? • Present the visualiza5on and cri5que to the class

  51. https://tinyurl.com/j2larrd • What is the visualiza5on about? • What informa5on is represented in the visualiza5on? And how? • What are the interac5ons used? • What ques5ons can we answer with the visualiza5on? • Do you like the visualiza5on? • Are there any improvements that can be made to the design?

  52. Assignment 1: Intro to Tableau Out today, due in two weeks • Use Tableau to visualize a simple dataset • Ask a meaningful ques5on about the data and answer the ques5on by crea5ng a good chart

  53. coming up Visual Integrity: How to not lie with visualiza5on

  54. Visual Integrity: How to not lie with visualization

  55. Visual Integrity: How to not lie with visualization The representa5on of numbers, as physically measured on the surface of the graphic itself, should be directly propor5onal to the quan55es represented

  56. New York Times 8/9/1978, via Edward Tufte and Andy Johnson (https://www.evl.uic.edu/aej/424/)

  57. Size of the effect in pixels/inches Lie Factor = Size of the effect in the data Effect size 7.8 (5.3-0.6) / 0.6 in graphics = = = 1 5.6 Effect size (27.5-18) / 18 0.5 lie factor in data New York Times 8/9/1978, via Edward Tufte and Andy Johnson (https://www.evl.uic.edu/aej/424/)

  58. Miles"per"gallon" 30" 25" 20" 15" 10" 5" 0" 1978" 1979" 1980" 1981" 1982" 1983" 1984" 1985"

  59. http://www.math.yorku.ca/SCS/Gallery/lie-factor.html Via Andy Johnson

  60. 70000" 60000" 50000" 40000" Male"professional" Physician"salary" 30000" 20000" 10000" 0" 1939" 41" 43" 45" 47" 49" 51" 53" 55" 57" 59" 61" 63" 65" 67" 69" 71" 73" 75"

Recommend


More recommend