Tue jack van wijk
- 1. 9/19/2016 1 Jack van Wijk Big Data Expo 2016, Utrecht 21-22 September, 2016 InformationVisualization Exploring the limits of what you can see in data JADS • Education, Research, Business • Three locations Mariënburg Convent, Den Bosch JADS www.jads.nl InformationVisualization • Overview • Trends • Examples InformationVisualization • The use of computer-supported, interactive, visual representations of abstract data to amplify cognition (Card et al., 1999) Data Visualization User
- 2. 9/19/2016 2 Why is my hard disk full? ? SequoiaView VanWijk andVan deWetering, 1999 Generalized treemaps • Idea: combine treemaps and business graphics • Many options Vliegen,VanWijk, andVan der Linden, 2006 Visualization high school data Cum Laude by MagnaView BIG DATA EXPO: STAND 01A BIG DATA EXPO: STAND 01A SequoiaView VanWijk andVan deWetering, 1999
- 3. 9/19/2016 3 Botanically inspired treevis What happens if we map abstract trees to botanical trees? Kleiberg et al., 2001 TreeViewBotanically inspired treevis Kleiberg,Van de Wetering, vanWijk, 2001 Botanically inspired treevis Kleiberg,Van de Wetering, vanWijk, 2001 Visualization of vessel traffic Willems et al., 2009 Visualization of vessel traffic Willems et al., 2009
- 4. 9/19/2016 4 InformationVisualization • The use of computer-supported, interactive, visual representations of abstract data to amplify cognition. (Card et al., 1999) Data Visualization User Infographics: - Static - Explanation - Made by data journalist - Viewed by lay audience Kentico.com Infographics vs InfoVis Infographics: - Static vs interactive - Explanation vs explorative - Made by data journalist vs domain expert - Viewed by lay audience vs domain expert Kentico.com Infographics vs InfoVis Multivariate Attributes Detail view Overview Selections Van den Elzen &VanWijk, IEEE InfoVis 2014 InformationVisualization • The use of computer-supported, interactive, visual representations of abstract data to amplify cognition • (Card et al., 1999) Data Visualization User The human visual system http://eofdreams.com
- 5. 9/19/2016 5 The human visual system http://vinceantonucci.com Translating data into pictures Position, width, height, colors encode six variables… Perception of symbols • How many red objects? Perception of symbols • How many red objects? Perception of symbols • How many circles?
- 6. 9/19/2016 6 Perception of symbols • How many circles? Perception of symbols • How many blue circles? Perception of symbols • How many blue circles? Limits to perception of symbols • Combinations of attributes cannot be perceived pre-attentively Color for encoding information • Translate data into colors • The human as light meter? www.weerdirect.nl Adelson checkerboard illusion
- 7. 9/19/2016 7 Adelson checkerboard illusion Use ColorBrewer for palettes Cynthia Brewer: http://colorbrewer2.org InformationVisualization • The use of computer-supported, interactive, visual representations of abstract data to amplify cognition (Card et al., 1999) Data Visualization User Data types simple hard multivariate data time series data hierarchical data networks text images video • Vary in complexity • One data set, many interpretations • Example Items with attributes age length sex name Multivariate data: tables 26 1.95 M 29 1.72 F Ivo Merel 57 1.85 MJack 57 1.68 FSimone age length sexname
- 8. 9/19/2016 8 Distribution per attribute 1.60 1.80 2.00 length 2 0 n 1 Events 1950 1960 1970 1980 1990 2000 Multivariate data: Parallel Coordinates Plot length 2.00 1.50 sex F M 50 age 50 10 Multivariate data: scatterplot 1.50 2.00 length 50 10 age Sets senior young F M Hierarchy senior young s y
- 9. 9/19/2016 9 similar age same sex Network One data set, many interpretations Abstract data: main types Multivariate visualization: scatterplot Tree visualization: tree diagram Graph visualization: node link diagram Abstract data: main types Multivariate visualization: scatterplot Tree visualization: tree diagram Graph visualization: node link diagram Challenge: What if we have thousands of data- items? Data size small (1-10) medium (1000) huge (> 106) business graphics infovis visual analytics Try to move to the left: • Filter, aggregate, statistics, machine learning, … without loosing essential information Anscombe’s quartet Francis Anscombe, 1973
- 10. 9/19/2016 10 Analysis of time-series data • Given: 10 minute measurements for one year • 52,560 measurements • How to visualize these? Analysis of time-series data • Given: 10 minute measurements for one year • 52,560 measurements • How to visualize these? • Cluster similar days • Use standard visualizations Analysis of time-series data Cluster & CalendarView, 1999 BaobabView Decision tree visualization,Stef van den Elzen, 2011 Big Data: D4D challenge Telecom data visualization,Stef van den Elzen, 2013
- 11. 9/19/2016 11 Abstract data: often a mix Multivariate visualization: scatterplot Tree visualization: tree diagram Graph visualization: node link diagram Hierarchy + network Holten, 2006
- 12. 9/19/2016 12 Spin-off: SynerScope • www.synerscope.com • Big Data Analytics • Transaction analysis, fraud detection BIG DATA EXPO: STAND 23 BIG DATA EXPO: STAND 23 InformationVisualization • The use of computer-supported, interactive, visual representations of abstract data to amplify cognition (Card et al., 1999) Check out www.jads.nl stand 01A stand 23 Thank you!