Separating Myth from Truth in Data Visualisation
14 likes5,820 views
This document outlines an agenda for a one-day workshop on data visualization and infographic design. The workshop aims to challenge existing thinking about creating and consuming visualizations, equip attendees with an appreciation of analytical and design choices, provide practice opportunities, and inspire attendees. The agenda covers fundamentals, the design methodology, data representation types, color theory, interactivity, and tools. Exercises are integrated throughout. The trainer, Andy Kirk, has extensive experience delivering visualization training globally.
![Visualising big data isn't a data problem, it's a summarisation
problem. You've only got so many pixels on the screen.
[Paraphrasing] Hadley Wickham
https://twitter.com/jsteeleeditor/status/434039532535562241
“Big Data” visualisation](https://image.slidesharecdn.com/andykirk-brighttalk-160811145859/85/Separating-Myth-from-Truth-in-Data-Visualisation-74-320.jpg)
Separating Myth from Truth in Data Visualisation
- 1. Separating Myth from Truth in Data Visualisation Andy Kirk www.visualisingdata.com @visualisingdata
- 2. Visualisation specialist: Editor of visualisingdata.com design consultant, trainer, lecturer, author, researcher... DATAVISUALISATION AND INFOGRAPHIC DESIGN Training Workshop 1-day ‘Introduction’ event The agenda is constantly refined to accommodate the most contemporary thinking and references. The final itinerary may be slightly different, and the precise times may vary, but here is an indicative outline of the typical 1-day agenda. 9:30 WELCOME: Introduction to today’s workshop 9:45 FUNDAMENTALS: What is data visualisation? 10:00 Exercise – Instinctive critical evaluations | Review 10:30 THE DATA VISUALISATION DESIGN METHODOLOGY 10:35 1. Understanding context andestablishing purpose 10:50 Exercise – Visualising ideas within problem contexts | Review 11:05 BREAK 11:15 2. Acquiring, preparing and familiarising with your data 3. Establishing editorial focus with your subject matter 11:35 Exercise – Exploring data and our editorial focus | Review 12:00 4. Conceiving your visualisation design Key principles of good visualisation design 12:15 LUNCH 13:05 Data representation, including a gallery of chart types and their roles 13:35 Exercise – Identifying data representation choices | Review 14:15 Colour theory and application Exercise – Forensic design assessments (1) 14:35 BREAK 14:50 Interactivity and animation features Exercise – Forensic design assessments (2) 15:10 Annotation and composition Exercise – Forensic design assessments (3) | Review 15:45 5. Construction, tools and technology 16:00 Exercise – Developing a data visualisation concept | Review 16:50 WRAP UP: The ‘8 Hats’ of visualisation design, workshop review 17:00 FINISH People attending these workshop sessions come from all backgrounds, organisation types and domain areas. You might be an analyst, statistician, or researcher looking to enhance your data capabilities. Perhaps you’re a skilled designer or developer looking to take your portfolio of work into a more data-driven direction. Maybe you do not personally get involved in the analysis or visualisation of data but coordinate others who do. You might be a frequent consumer of visualisation and infographic work looking to increase the sophistication of how you read, interpret and evaluate the effectiveness of such designs. These events are intended for all audience types and levels. The most critical attribute is that you have an inherent curiosity, an instinct for discovering and sharing insights from data, and an interest in approaching your data visualisation and infographic work with a fresh perspective. You should have a willingness to contribute to discussions during exercise activities and do so in a respectful and constructive manner. To view a selection of testimonials from previous workshop participants visit www.visualisingdata.com and click on Training. The ‘Introduction to Data Visualisation and Infographic Design’ 1-day workshops aim to provide delegates with an accessible and comprehensive introduction to data visualisation and infographic design. The focus for this training is the craft of this discipline, helping delegates to know what to think, when to think about and how to resolve all the analytical and design decisions involved in any data-driven challenge. There are four over-riding objectives for these workshops: To challenge your existing thinking about creating and consuming visualisation works, refining the clarity of your convictions about effective visualisation design. To equip you with an appreciation of all the analytical and design choices available across the creative workflow: the options that exist and what decisions to take. Tho provide an opportunity to practice by undertaking focused activities at each stage of the creative workflow, applying and cementing the learning at each stage. To inspire you by broadening your visual vocabulary, by exposing you to the latest techniques and contemporary resources, and by giving you the confidence to enhance your data visualisation capabilities. Andy Kirk is a UK-based data visualisation specialist: A design consultant, training provider, lecturer, author, blog editor, speaker, and researcher. Since launching his freelance career in late 2011, he has delivered nearly 150 public and private training events in 14 countries across five continents. Recent clients include PepsiCo, Standard Chartered, Cisco, Heineken, and CERN. Visit www.visualisingdata.com and click on Training to see a map and profile of his past training events. Andy’s teaching activities extend to a visiting lecturer position at the Maryland Institute College of Art (MICA) in the US where he teachers a module on the Information Visualisation (MPS) Masters programme. The workshop session is structured around a proven design methodology where we will build up, stage by stage, a detailed understanding of all the different aspects of decision-making that goes into all data visualisation or infographic design work. Whilst the event is described as an ‘Introduction’ this does not mean it is pitched at a ‘basic’ level: Data visualisation and infographic design teaching is really framed by the time available - and the breadth/depth achievable - rather than any distinction between content progressing from beginner towards advanced topics. The content will be delivered through a blend of teaching, discussion, and group practice. The practical exercises vary in nature from evaluating works, conceiving ideas, identifying best-fit solutions and exploring data. The most technical of the exercises will involve looking at data in Excel. There will also be some sketching tasks but note these are not a test of artistic capability, rather conceptual thinking. These workshops are designed to be technology neutral, they are not focused on or based around any specific tool or programme. However, there will be a section providing an overview of some of the most essential and common visualisation tools currently in the market. All materials will be issued digitally on USB memory sticks containing all training content, exercise files and a range of useful references. Attendees are encouraged to bring laptops to use as a workspace for the session. The only software requirements are Excel, a browser and a pdf reader. No other technical prerequisites exist. TRAINING OBJECTIVES THE WORKSHOP WHO SHOULD ATTEND? TYPICALAGENDA TRAINER PROFILE REGISTER NOW! Visit www.visualisingdata.com and click on Training to find an event nearyou
- 3. The representation and presentation of data to facilitate understanding What is data visualisation?
- 4. In data visualisation we often hear about the always and the nevers but, invariably, it is far more nuanced and requires more sophisticated convictions. Which things are the always and which are the mostlys? Which things are the nevers and which are the rarelys? “Separating Myth from Truth in Data Visualisation”
- 6. ______ charts are evil
- 7. “Pie charts are evil...”
- 8. Robert Kosara: “In defense of pie charts” https://eagereyes.org/criticism/in-defense-of-pie-charts
- 9. Maximum 2-3 parts of a whole, start at the vertical
- 10. http://www.bricksetforum.com/discussion/21529/colors-how-theyve-changed Value of pie charts for exploring data (vs. communicating)?
- 11. “Word clouds are evil...”
- 12. Google search: “Marketing Infographics” “Infographics are evil...”
- 14. BAD PIE CHARTS ARE BAD CHARTS BAD BAR CHARTS ARE BAD CHARTS BAD WORD CLOUDS ARE BAD CHARTS BAD RADAR CHARTS ARE BAD CHARTS Repeat after me...
- 15. No chart is evil: They play different roles & all have limitations
- 16. 3D
- 17. http://blog.visual.ly/2ds-company-3ds-a-crowd/ Distorting decoration: Don’t create 3D out of 2D data
- 19. Image from https://twitter.com/aquigley/status/739752961920933888 When is 3D legitimate?: When consuming in 3D...
- 20. http://www.nytimes.com/interactive/2015/03/19/upshot/3d-yield-curve-economic-growth.html?smid=pl-share&_r=1&abt=0002&abg=0 When is 3D legitimate?: When representing 3D data...
- 21. http://www.nytimes.com/interactive/2015/03/19/upshot/3d-yield-curve-economic-growth.html?smid=pl-share&_r=1&abt=0002&abg=0 When is 3D legitimate?: ...and when offering a series of 2D views
- 22. Truncated axes
- 23. Truncated axes: Bar charts
- 26. Truncated axes: Area charts
- 28. $0.8M out of $7.5M = 10.7% Length of presented bar progress = 24.6% Size miscalculations
- 29. If the value of A is twice as big as the value of B...
- 30. Data:Ink ratio & “Chart Junk”
- 31. Data:Ink Ratio: Too brutal, even for the era of print-only http://www.perceptualedge.com/articles/visual_business_intelligence/sometimes_we_must_raise_our_voices.pdf
- 32. http://seeingdata.demo.cleverfranke.com/census/ Data:Ink Ratio: “Ink” creates unity, contrast, organisation
- 33. Data:Ink Ratio: “Ink” creates unity, contrast, organisation
- 34. https://paragraft.wordpress.com/2008/06/03/the-chart-junk-of-steve-jobs/ Chart Junk: Incongruent or obstructive
- 35. https://www.statista.com/chart/4101/where-is-pensioner-poverty-the-most-prevalent/ | https://twitter.com/olafcramme/status/523397542172950528 | http://exceluser.com/blog/1152/oh-no-chart-junk-from-the-wall-street-journal.html Chart Junk: Incongruent or obstructive
- 36. Chart junk? Congruent and unobstructive http://www.wsj.com/articles/SB12147335600370333763904581058081668712042 | http://www.bloomberg.com/news/features/2015-03-03/junk-food-s-last-stand-the-pizza-lobby-is-not-backing-down
- 37. Rainbow colour schemes for showing quantitative values
- 38. http://eagereyes.org/basics/rainbow-color-map Multi-hue colour scheme: Magnitude & order?
- 39. http://www.nzz.ch/aktuell/inland-sommerserie-schweizer-karten-interaktiv/vergleich-strom-versorger-1.18120907 Convergent colour scheme: Magnitude & order
- 40. Divergent colour scheme: Magnitude & order http://www.randalolson.com/2015/09/05/visualizing-indego-bike-share-usage-patterns-in-philadelphia-part-2/
- 41. Red & green (should they ever be seen?)
- 42. Impact of colour blindness (deuteranopia)
- 43. #a6611a #7b3294 #d01c8b #018571 #008837 #4dac26 #d7191c #2c7bb6 To mean ’GOOD’ To mean ’BAD’ Alternative colours to replace the default green and red
- 44. Maybe nobody in your audience is colour-blind?
- 46. “Data visualisation is all about storytelling”
- 47. Stories can be seen through temporal charts http://www.nytimes.com/interactive/2016/us/elections/polls.html?_r=0
- 48. Stories can be seen through temporal charts
- 49. Stories can emerge from temporal presentations
- 50. Stories can emerge from temporal presentations
- 51. http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0123507 Stories can emerge from temporal presentations
- 52. https://www.ted.com/talks/hans_rosling_shows_the_best_stats_you_ve_ever_seen?language=en Stories can emerge through verbal annotation (chart as a prop)
- 53. Difference between storytelling and storyforming The stories we form in our mind through interpretation
- 54. Difference between storytelling and storyforming The stories we form in our mind through interpretation
- 55. “Data visualisation is about perceptual accuracy”
- 56. Ranking of elementary perceptual tasks From Cleveland & McGill, ‘Graphical Perception: Theory, experimentation and application to the development of graphical methods’”, 1984
- 57. B=5 A=10 Length Area A=10 B=5 Ranking of elementary perceptual tasks Discernibility of attributes: Quantitative data
- 58. “The ordering does not result in a precise prescription for displaying data but rather is a framework within which to work.” Cleveland & McGill Ranking of elementary perceptual tasks From Cleveland & McGill, ‘Graphical Perception: Theory, experimentation and application to the development of graphical methods’”, 1984
- 59. What’s the right level of discernibility?
- 60. What is the shape of the data you are trying to encode?
- 61. “Data visualisation is about simplicity”
- 64. Simple subject, shown as clearly as possible
- 65. https://twitter.com/theboysmithy/status/705323516711804928 Complex subject, shown as clearly as possible
- 67. “Data visualisation is about immediate understanding”
- 68. http://oecdbetterlifeindex.org/ Some things can be quickly consumed...
- 70. Some visualisation techniques are familiar...
- 72. Learnability: Everything is new, once
- 73. “Visualising ‘Big Data’ requires special techniques”
- 74. Visualising big data isn't a data problem, it's a summarisation problem. You've only got so many pixels on the screen. [Paraphrasing] Hadley Wickham https://twitter.com/jsteeleeditor/status/434039532535562241 “Big Data” visualisation
- 77. CLOSING REMARKS
- 78. In the absence of perfect, optimisation is the pursuit: Maximum effectiveness through maximum efficiency Data visualisation is always about compromise
- 79. To make the best decisions you need to be familiar with all your options and aware of the things that will influence your choices. Data visualisation is always about good decisions THINGS YOU COULD DO THINGS YOU WILL DO
- 80. BOOK! Data Visualisation: A Handbook for Data Driven Design
- 84. Separating Myth from Truth in Data Visualisation Andy Kirk www.visualisingdata.com @visualisingdata