Show me the Numbers: Designing Your Own Data Visualizations

Data visualization basics at the PEPFAR Applied Learning Summit

https://achafetz.github.io/pals_dataviz/

Transcript

1. SHOW ME THE NUMBERS Designing your own data visualizations September 2017
3. What is a tool?
4. How do we design an effective tool?
5. How do we design an effective tool? Data Viz Principles Program Context Software Skills Structured Dataset Guiding Question
6. What can we learn from a car dashboard?
7. How many ways can you represent…. 75 37 and
8. A few examples… http://blog.visual.ly/45-ways-to-communicate-two-quantities/
10. Global Single IM Broad Focus Narrow Focus PEPFAR Executives TWG Leads SI Advisers Partners Country Teams Agency Leadership Defining an audience Scott Jackson, PWC
11. Explore Analyze Communicate Influence A good visualization allows you to… Chafetz, Essam, Hughes, Johnson (2016). “Fundamentals of Data Analysis & Visualization Training”
12. Explore Analyze Communicate Influence Extract Clean Validate ….but first you need to start by setting up the data
13. Extract Clean Validate Explore Analyze Communicate Influence Tool Building Process
14. Where do you look for data?
16. We are interested in tracking the progress of the number of people on treatment by age group against the COP target set in PEPFARlandia. • Where do we go to access and extract the data? • What information do we need to pull the data?
17. Extract Clean Validate Explore Analyze Communicate Influence Tool Building Process
18. Structure Integration Subset
19. Differing structures even within PEPFAR data systems
20. Extract Clean Validate Explore Analyze Communicate Influence Tool Building Process
21. 80% How complete is the dataset?
22. 775 777 745 75 779 785 786 Total Numerator Aggregated Age/Sex Results Age/Sex Aggregated/Result Age/Sex/Result ServiceDeliveryPoint ServiceDeliveryPoint/Result PEPFARlandia Completeness Check FY 2016 Q2 HTS_TST thousands of people
23. Validate the completeness of TX_CURR for DSD for the current FY in PEPFARlandia. • Where do we go to access and extract/view the data? • What information do we need to pull the data? • What disaggregates do we need to look at?
24. How valid is the tool? Panorama Tool Investigate Adjust Tool Re-Validate
25. Extract Clean Validate Explore Analyze Communicate Influence Tool Building Process
27. We are interested in tracking the progress PEPFARlandia is making. • Using the handouts containing PEPFARlandia indicator data, explore the data by making some visualizations. • What story are you trying to tell with the visualization?
28. Extract Clean Validate Explore Analyze Communicate Influence Tool Building Process
29. Visual Analysis to Find Stories Andy Kirk, Visualization Workflow o Categorical comparison and proportions o Ranking: big, medium, small o Measurements/values: absolutes o Range and distribution o Context: Targets, forecasts, averages o Hierarchical relationships COMPARISONS o Up and down vs flat? o Linear vs exponential o Steady vs fluctuating o Seasonal vs random o Rate of change vs steepness TRENDS o Outliers o Intersections o Correlations o Connections o Clusters o Associations o Gaps RELATIONSHIPS
30. We are interested in tracking the progress PEPFARlandia is making. • How were you going about analyzing the data your visualizations? • How could you improve upon your visualization?
31. Tools for Analysis
32. Extract Clean Validate Explore Analyze Communicate Influence Tool Building Process
33. Static Dynamic Explanatory Exploratory Form vs. Function
34. Static/Explanatory https://www.washingtonpost.com/graphics/local/metro-closures/
35. Dynamic/Explanatory http://flowingdata.com/2016/05/10/who-is-older-and-younger-than-you/k
36. Static/Exploratory http://www.wsj.com/articles/u-s-employers-add-271-000-jobs-1446816737
37. Dynamic/Exploratory http://www.nytimes.com/interactive/2015/11/24/upshot/thanksgiving-flight-patterns.html
38. Consistent color scheme Consistent font family, vary size and emphasis t T Adjust size Simplify Logical Order Integrate text and legend MWI TNZ Start with a sketch Key Principles Clear purpose
39. We are interested in tracking the progress PEPFARlandia is making. • How could you improve upon your visualization? • Add in some features mentioned on the last slide around key principles, starting with a clear title
40. Extract Clean Validate Explore Analyze Communicate Influence Tool Building Process
42. Extract Clean Validate Explore Analyze Communicate Influence Tool Building Process
43. Let’s critique some data visualizations
63. • Prepared for the ICPI/CIT training, “Data Use Training” (DC), October 2016 (adapted from June 2016 version) • Presentation material adopted from J. Schwabish (2014). “A Visualization Mapping: Form and Function” (http://policyviz.com/a-visualization-mapping-form-and-function/), S.Ortiz (2012). “45 Ways to Communicate Two Quantities, (http://blog.visual.ly/45-ways-to-communicate- two-quantities), A. Kirk (2013). Visualization WorkflowFinding Stories and Telling Stories (http://www.slideshare.net/visualisingdata/andy-kirks-facebook-talk) and Chafetz, Essam, Hughes, Johnson (2016). “Fundamentials of Data Analysis & Visualization Training” (http://geocenter.github.io/StataTraining ) • Image Sources • Youtube [Dashboard], Sears Auto Center • Unsplash, [Lake and Mountains], Justin Luebke • Extreme Presentation, “Chart Suggestions” • Icons downloaded from the Noun Project and designed by various artists: The Noun Project, [map], Ivan Colic; “Microscope”, lastspark, “Graph”, gira Park; “Influencer”, Adam Beasley; “Pickaxe”, Creative Stall, “Detergent”, Megan Mitchell, “Washing Machine”, National Parks Service, “Clipboard”, matthew hall, Phil Goodwin; [country shape], Anna Gajowiec, “Traffic Cone”, Vicons Design, “Structure”, Alexandr Cherkinshy, “simplify”, Chameleon Design, “Analytics, Syafiqa Fickle; “Pencil”, David; “Target, Franck Junker, “Mouse”, John Testa, “Wrench”, Bluetip Desig, “Power Drill” Maksim Karalevich Notes and Attribution