Designing Data Visualizations to Strengthen Health Systems
7 likes1,469 views
The document discusses the importance of effective data visualization in strengthening health systems, emphasizing the necessity to convey information accurately and compellingly. It covers key concepts, historical examples, audience personas, and best practices for designing data visualizations while highlighting common pitfalls to avoid. The authors, Amanda Makulec and Jeff Knezovich, aim to inspire and bolster the capacity for impactful data storytelling through visual means.
![“The two optic nerves [in the
eyes] are sending what we
now know are 20 megabits a
second of information back to
the brain.”
- Edward Tufte](https://image.slidesharecdn.com/datavizdesignworkshop-hsr2016-final-161117002453/85/Designing-Data-Visualizations-to-Strengthen-Health-Systems-43-320.jpg)
Designing Data Visualizations to Strengthen Health Systems
- 1. Designing Data Visualizations to Strengthen Health Systems Amanda Makulec Visual Analytics Advisor,JSI Jeff Knezovich Director,Quaternary Consulting Health Systems Research Symposium 15 November 2016 |Vancouver,BC,Canada
- 2. Amanda Makulec, MPH Visual Analytics Advisor John Snow Inc. Jeff Knezovich Director Quaternary Consulting
- 5. conceptual data driven declarative exploratory idea illustration everyday data viz idea generation visual discovery Matrix credit: Harvard Business Review’s Good Charts
- 6. Data Journalism Learn more on how to craft effective data journalism at http://datajournalismhandbook.org/
- 8. Where did viz start?
- 10. What is data visualisation? • A way of visually conveying information – often quantitative in nature – in an accurate, compelling format. • Usually makes relationships more apparent (e.g. by clustering, color coding and by putting items in scale). • Can be static or interactive.
- 11. Data visualisation, or information visualisation? 1001000 1100101 1101100 1101100 1101111
- 12. Data visualisation, or information visualisation? 72 101 108 108 111
- 13. Data visualisation, or information visualisation?
- 14. Brief history
- 15. Are data visualisations new? • William Playfair – 1786: line graph and bar chart of economic data – 1801: pie chart and circle graph
- 16. • Florence Nightingale – 1858 polar area diagram Are data visualisations new?
- 17. Are data visualisations new? • John Snow – 1854: Mapping deaths from a Cholera outbreak in central London
- 19. Each persona represents a significant portion of people in the real world and enables the designer to focus on a manageable and memorable cast of characters, instead of focusing on thousands of individuals.” persona, n. “A persona is depicted as a specific person but is not a real individual; rather, it is synthesized from observations of many people. From: https://www.smashingmagazine.com/201 4/08/a-closer-look-at-personas-part-1/
- 20. Interests?
- 21. Motivations?
- 22. Action?
- 23. steps to develop simple data viz audience personas
- 24. 1. Pick your stakeholder groups
- 25. 2. Identify key personas within a stakeholder group
- 26. sample template
- 27. 3. Map personas by dimension.
- 28. Factual, analytical Feeling, intuitive Personal gain System gains External recognition Internal reward Simplicity Complexity Complacent Driver of change Debilitated by chaos Thrives on chaos Team oriented Individual/loaner Problem solver Defeatist Black and white Compromise Receptive Rigid Team leader Team member Carrot eater Stick driven Short term focus Long term vision Self accountable Cheater Internal motive External motive Technical Political Needs clarification Self-motivated Empowerer Underminer Works better in group Works better alone Data driven Story motivated Values independence Values collaboration Head Heart Personas on Continuums Much like the Myers-Briggs personality scale, personas can be ranked along continuums of characteristics that may impact their use of data for decisionmaking.Some examples of different “poles” identified in workshops are below.
- 29. Ministry of Health Rachel The Technocrat Pascal The Politician for example segmented to
- 30. Wants stories Wants numbers Motivated by passion Motivated by money Act based on feelings Act based on data Resilient problem solver Frustrated bureaucrat Champion for your issue Oppose your issue mapping by dimension
- 31. Wants stories Wants numbers Motivated by passion Motivated by money Act based on feelings Act based on data Resilient problem solver Frustrated bureaucrat Champion for your issue No knowledge of your issue mapping by dimension Rachel the Technocrat Pascal the Politician
- 33. Not just job titles stakeholder groups organization names Focus on the human side of your data viz audience.
- 34. Let’s look at some visualizations.
- 35. WTF is wrong with the following visualisations?
- 36. WTF is wrong with: Image from Patients Association: http://www.patients-association.org.uk/reports/waiting-times-report-feeling-wait/
- 37. WTF is wrong with: Image from WTFviz: http://visual.ly/buzzing-trends-real-estate-market-bangalore
- 38. WTF is wrong with: Image from WTFviz: http://visual.ly/beyond-facebook-marketing-new-generation
- 39. WTF is wrong with: Image from WTFviz: http://viz.wtf/image/110276700184
- 40. WTF is wrong with: Image from WTFviz: https://twitter.com/BofA_News/status/558242696415166464
- 41. On Think Tanks Data Visualisation Competition Objectives: Inspire Strengthen capacity Encourage
- 43. “The two optic nerves [in the eyes] are sending what we now know are 20 megabits a second of information back to the brain.” - Edward Tufte
- 44. System 1 vs System 2 thinking 17 x 24 = ? Example from Graham Odds
- 45. Conscious v sub-conscious bandwidth 0 2 4 6 8 10 Taste Auditory Olfactory Tactile Visual Sub-conscious (millions of bits per second) 0 10 20 30 40 Conscious (bits per second) Adapted from: Tor Norretranders' The User Illusion
- 46. make data sticky
- 47. Gestalt design: taking advantage of sub-concsious processing
- 48. Gestalt design principles: Implications for data visualisation Adapted from Alberto Cairo
- 49. Gestalt design principles: Semiotics and iconography ☺
- 50. Who makes a good data visualisation?
- 51. Who makes a good visualisation? Communication Research Design Technology From: https://onthinktanks.org/art icles/visualising-data-both-a- science-and-an-art/
- 52. • Data literacy – merging and tidying datasets • Statistical competencies – mean v median, ordinal versus scalar • Research methods - sampling • Research context Research
- 53. • For data collection – e.g. web scrapers • For data storage – e.g. database and SQL • For data manipulation – e.g. SPSS, R • For data visualisation – e.g. coding, like jQuery, HTML5 Technology
- 54. • Appropriate visualisation types • Chart fundamentals • Colour and form Design
- 55. • Understanding of users’ needs and perceptions • Finding and refining messages • Telling stories Communication
- 56. What makes a good data visualisation?
- 57. What makes a good visualisation? Interesting Function Form Integrity (McCandless, 2012)
- 58. Function Source: Gregor Aisch - http://slidesha.re/1jTg5Eq
- 59. Source: Adapted from from The Wall Street Journal guide to information graphics Form
- 60. Integrity
- 61. From: XKCD CONSUMERS OF BOILED EGGS Interestingness
- 62. What do you think are key elements of effective data visualisation?
- 63. preattentive attributes strategic chart selection
- 64. “ t h i n k i n g w i t h y o u r l i z a r d b r a i n ” preattentive attributes
- 65. Paired ColumnColumn Bar Paired Bar Stacked Bar Stacked Column Slope compare categories
- 66. Histogram Box and Whiskers Confidence Interval distribution
- 68. Line Stacked Area SparkLines Dot Plot time series 0 20 40 60 80
- 69. why dots? 0 20 40 60 80 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8 Year 1 Year 5 0 20 40 60 80 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8
- 74. Tree Map 0% 20% 40% 60% 80% 100% Icon Matrix part-to-whole better 839 294 145 50 Small Multiples Chart Horizontal Stacked Bar
- 76. icon matrix
- 77. pick your icon
- 79. for social share & infographics
- 80. View & order a copy at http://policyviz.com/graphic-continuum/ chart type quick reference
- 81. Common Pitfalls
- 82. Things to avoid when designing your visualizations. Many courtesy of Gregor Aisch (http://slidesha.re/1jTg5Eq)
- 84. Going colour crazy 0 1 2 3 4 5 6 7 8 9 10 Category 1 Category 2 Category 3 Category 4 0 1 2 3 4 5 6 7 8 9 10 Category 1 Category 2 Category 3 Category 4 vs
- 85. Quick colour aside Analogous (similar colours) monochromatic complementary From: Data visualisation: a practical guide to producingeffective visualisations for research communication http://resyst.lshtm.ac.uk/resources/data-visualisation-practical-guide-producing-effective-visualisations-research
- 86. Not remembering the objective Team 1 Team 2 Person A Person B Person C Person D Person E Person F Person G Person H vs
- 87. Not thinking about order 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 A B C D 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 D A C B vs
- 88. Not scaling to zero 11.5 12 12.5 13 13.5 14 14.5 15 Category 1 Category 2 Category 3 Category 4 0 2 4 6 8 10 12 14 16 Category 1 Category 2 Category 3 Category 4 vs
- 89. Not scaling to zero (exceptions) vs
- 90. Not labelling directly Series 1 Series 2 Series 3 0 1 2 3 4 5 6 1990 1995 2000 2005 0 1 2 3 4 5 6 1990 1995 2000 2005 Series 1 Series 2 Series 3 vs
- 92. Non-descriptive titles Youth Unemployment Rates in Europe
- 93. Non-descriptive titles Youth Unemployment on Historical High
- 94. Non-descriptive titles Youth Unemployment Divides Europe
- 95. Non-descriptive titles Youth Unemployment Divides Europe Seasonally adjusted unemployment rates of under 25s
- 96. Seven deadly sins of data visualisation https://onthinktanks.org/articles/on-datavis-judging-jeff-knezovichs-advice/ From: https://onthinktanks.org/articles/on- datavis-judging-jeff-knezovichs-advice/
- 101. Ambiguity
- 102. Over- complication
- 103. Forgetting the point
- 105. Designing in Excel
- 107. Simple LESS IS MORE white space is nice
- 108. three things to improve every chart
- 109. 1: declutter
- 111. 8 14.9 11.7 38.9 19.2 16.6 20.4 26.2 0 5 10 15 20 25 30 35 40 45 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8 Year 1 Year 1
- 112. 0 10 20 30 40 50 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8
- 115. 2: color sparingly
- 116. 0 10 20 30 40 50 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8
- 117. 0 10 20 30 40 50 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8
- 118. 0 1 2 3 4 5 6 Category 1 Category 2 Category 3 Category 4 Series 1 Series 2 Series 3 Super special brand-compliant graph! Inspired by Cole Naussbaumer at Stanford’s Data on Purpose, February 2016
- 119. 4.3 2.5 3.5 4.5 Category 1 Category 2 Category 3 Category 4 Series 1 Series 2 Series 3 Easier to see a key data series graph! Inspired by Cole Naussbaumer at Stanford’s Data on Purpose, February 2016
- 120. be mindful
- 121. custom styles
- 122. 3: purposeful title
- 123. 0 10 20 30 40 50 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8 Quality of Care Score by Facility Ghana, 2015
- 124. 0 10 20 30 40 50 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8 Facility 4 showed the highest quality of care. Despite scoring highest, its overall score was below 50%, indicating there is work to be done to improve quality of care.
- 125. 38.9 Facility 1 Facility 2 Facility 3 Facility 4 Facility 5 Facility 6 Facility 7 Facility 8 Facility 4 showed the highest quality of care. Despite scoring highest, its overall score was below 50%, indicating there is work to be done to improve quality of care.
- 126. chart templates
- 127. favorite things Graphic Continuum Dot plots Icon matrix Canva Piktochart Noun Project Visage DataViz 101 Pexels Slidedocs Google Sheets High Charts
- 128. Inspiration Skill building data viz recs
- 129. duarte.com // storycorps.org // skillshare.com storytelling recs