Data visualization
Download as PPTX, PDF0 likes276 views
This document provides an overview of different types of data visualization techniques including tables, charts, dashboards, key performance indicators (KPIs), sparklines, word clouds, advanced charts, storytelling tools, and exploration tools. It discusses best practices for using the right chart type based on the data, avoiding misleading pie charts, considering colorblind viewers, using 3D only in specific cases, applying Gestalt principles of perception, and choosing visualizations based on the audience. The document emphasizes that the right chart depends highly on the context and story being told.
Data visualization
- 1. F A R E W E L L T O U R DUTCH DATA DUDE JEROEN TER HEERDT Technical Evangelist @jeroenterheerdt http://www.dutchdatadude.com Data visualization #farewelltour #dutchdatadude
- 8. Data Artist
- 12. Tables Most basic form of data visualization Around since 3000 BC, perfected by Romans Useful for lists
- 13. Charts Around since 14th century Easy to create, easy to grasp Started with bar- and line charts
- 14. Dashboards Invented around 1970, really took of in 1990 with the web At-a-glance view of data, obscuring details Managers love these
- 15. KPIs KPI = Key Performance Indicator Invented around 1990 Great for targets / budgets Can include trend as well
- 16. Sparklines Smaller charts No axis, no labels Only bare necessities Frequently used in combination with tables Also called condensed charts
- 17. Word clouds Easy way to show terms and their relative occurrence or relevance
- 18. User- oriented Group of chart types that aim to focus on how the user consumes in the information
- 19. Advanced Powerful charts, can be harder to understand The sky is the limit
- 21. Story telling Story telling tools Power Map SketchStory
- 23. Exploration Q&A
- 27. Pie charts Typically used for part-of-a-whole visualization Very popular with management A pie is for eating only, not for data visualization
- 28. Pie charts Better solution for part-of-a-whole visualization: bar / column chart
- 29. Color usage Think about the color blind
- 30. 3D 3D is just a dimension wasted If you need to use 3D to spice up your information display, it is not worth displaying
- 32. Choose the right chart type There is a lot to this, but here are some ideas Bullet chart for KPI Treemap Small multiples / Trellis chart Scatter plot
- 33. Choose the right chart type There is a lot to this, but here are some ideas
- 34. Audience What level of information does the reader need? What is the story I want my reader to understand? What message do I want to bring across? Social context. Meaning of icons / colors.
- 35. Best practices summary Rule #1: A pie is for eating only; never use pie charts Rule #2: Think about the color blind; use the right color scheme Rule #3: 3D is just a dimension wasted, only very specific scenarios it is useful (think maps) Rule #4: Gestalt principles of perception Rule #5: Choose the right chart type; hard to do, but worth it. Rule #6: Be mindful of the context of the audience
Editor's Notes
- #5: Deze drie functies heb je nodig. Data Wrangler / Scientist / Sculptor (Artist). Het is aan een ieder om zijn rol te kiezen, daarbij ondersteund door technologie en platform die maakt dat je op je best bent.
- #6: Deze drie functies heb je nodig. Data Wrangler / Scientist / Sculptor (Artist). Het is aan een ieder om zijn rol te kiezen, daarbij ondersteund door technologie en platform die maakt dat je op je best bent.
- #13: Clay tablet
- #14: Clay tablet
- #15: Clay tablet
- #20: Vergeet SAS, SPSS. R is de standard nu.
- #21: Gemaakt met Excel
- #23: demo
- #24: Met handschrift en spraak demo
- #25: demo
- #28: Imagine a clock (#1): this is clearly five o’clock, right? Same clock again (#2): what is the time now? Harder right? Maybe 3:55? Let’s add numbers (#3): now it is easier, because we gave you some help. This is normal for a clock Clock #4: extra hand (for seconds): makes it even harder to read. Which is bigger? A,B or C? Clock #5: add a fourth hand, even harder to read Clock #6: we could add data labels to make it easier to use, but this is a workaround for a core problem: in our brains we just cannot easily judge the slice that is determined by two lines from a center. Better use 100% bar chart for part of a whole, since it makes use of only vertical and horizontal lines, not angled lines.
- #29: Better solution for part-of-a-whole visualization: (100%) bar/column chart. Click: only use for a pie chart…
- #33: Bullet charts work because we immediately in our brains recognize the cross shape. Scatterplot demo.
- #34: Bullet charts work because we immediately in our brains recognize the cross shape. Scatterplot demo.