Presentation is loading. Please wait.

Presentation is loading. Please wait.

1 18 April 2007 vizNET-LEEDS-PRES-0002-070418 A Rough Guide to Data Visualization – Part 2 VizNET 2007 Annual Event Ken Brodlie School of Computing University.

Similar presentations


Presentation on theme: "1 18 April 2007 vizNET-LEEDS-PRES-0002-070418 A Rough Guide to Data Visualization – Part 2 VizNET 2007 Annual Event Ken Brodlie School of Computing University."— Presentation transcript:

1 1 18 April 2007 vizNET-LEEDS-PRES-0002-070418 A Rough Guide to Data Visualization – Part 2 VizNET 2007 Annual Event Ken Brodlie School of Computing University of Leeds

2 2 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Outline of Session  Visualizing Tables of Information  Scatter plot matrices  Parallel coordinates  Glyphs  Visualizing Structures of Information  Interacting with a Visualization  Focus and context  Visualizing Time

3 3 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Visualizing Tables of Information Multivariate datasets

4 4 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Data Tables  Multivariate datasets can be expressed as a data table  Each entry in table is an observation  An observation consists of values of a set of variables, or variates ABC 1.. 2 variables observations

5 5 18 April 2007 vizNET-LEEDS-PRES-0002-070418 What are these?

6 6 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Multivariate Visualization  Example of iris data set  150 observations of 4 variables (length, width of petal and sepal)  Check wikipedia for explanations of petals & sepals  Techniques aim to display relationships between variables – the analytical task Challenge in visualization is to design the visualization to match the analytical task

7 7 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Scatter Plot  For two variates, we have already met the scatter plot technique  It is useful for showing what happens to one variable as another changes…

8 8 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Scatter Plot Matrix – Iris Data Set For a data table of N > 2 variables, we can look at all pairs of variables in a matrix of scatter plots. What can you observe?

9 9 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Parallel Coordinates – Iris Data  One axis per variable  Scaled to range  Each observation is plotted as a ‘polyline’  Correlations may start to appear

10 10 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Parallel Coordinates  Concept due to Alfred Inselberg  Conceived the idea as a research student in 1959…  … idea gradually refined over next 40 years http://www.math.tau.ac.il/~aiisreal/

11 11 18 April 2007 vizNET-LEEDS-PRES-0002-070418 The Screen Space Problem  All techniques, sooner or later, run out of screen space  Parallel co-ordinates  Usable for up to 150 variates  Unworkable greater than 250 variates Remote sensing: (5 variates, 16,384 observations)

12 12 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Brushing as a Solution  Brushing selects a restricted range of one or more variables  Selection then highlighted Note outlier

13 13 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Glyph Techniques – Star Plots  Each observation represented as a ‘star’  Each spike represents a variable  Length of spike indicates the value

14 14 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Glyph Techniques – Star Plots – Iris Data  Each observation represented as a ‘star’  Each spike represents a variable  Length of spike indicates the value Iris data set

15 15 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Multivariate Visualization Techniques  Software:  Xmdvtool Matthew Ward n Techniques designed for any number of variables –Scatter plot matrices –Parallel co-ordinates –Glyph techniques http://davis.wpi.edu/~xmdv

16 16 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Visualizing Structures of Information

17 17 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Examples of Networks of Information My Window filestore Automobile web site - visualizing links

18 18 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Suppose user has three subdirectories: A, B and C First partition in X according to total size of each sub- directory ABC Tree Map of Filestore Note we are seeing the relative sizes of directories

19 19 18 April 2007 vizNET-LEEDS-PRES-0002-070418 ABC Then within each subdirectory, we can partition in Y by the size of individual files, or further subdirectories Tree Map of Filestore

20 20 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Treemap Example Usenet news groups For history of treemaps see: www.cs.umd.edu/ hcil/treemap-history Developed over many years by Ben Schneiderman and colleagues

21 21 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Tree Maps – A Filesystem Download SequoiaView from: http://www.win.tue.nl/sequoiaview/ -Developed by Jarke van Wijk and colleagues Can be hard to decipher so..

22 22 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Visualizing Structure of Web Pages  Star trees have been successfully used to visualize web site structures  Place home page in centre  with linked pages connected by hyperbolic arcs  further arcs link to further links Escher woodcut….. inspired:

23 23 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Star Trees Automobiles web site Home page in centre Click on link you want...

24 24 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Star Trees Auto History moves to centre of screen Click on next link... And so on

25 25 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Star Trees  Commercially available from Inxight  Spun out of Xerox PARC  www.inxight.com www.inxight.com

26 26 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Visualizing Web Searches www.kartoo.co.uk

27 27 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Interacting with Visualizations

28 28 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Interaction  Major difference between paper and computer-based visualization is ability to interact with picture and alter the presentation of the data  We have already seen brushing and here we look at the idea of focus and context  A recurring problem in visualization is lack of screen real estate  Want to achieve:  Focus: to see detail of immediate interest  Context: to see the overall picture  Want to do this interactively…

29 29 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Bifocal Display  Magnify focus, demagnify context  A possible application is in map reading  This London underground application was developed by Marcelo Cohen in Leeds  Zoom / pan  Focus /context

30 30 18 April 2007 vizNET-LEEDS-PRES-0002-070418 What is the Bifocal Display Doing?  Transforming the information space to the display space  Visual transfer functions Information space Display Space Normal display Information space Display Space Bifocal display context focus

31 31 18 April 2007 vizNET-LEEDS-PRES-0002-070418 The Perspective Wall 2D layout wrapped around a 3D structure Space utilisation: -detail on centre panel 3x size of equivalent flat wall fitting field of view

32 32 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Multiple Views  Multiple views can also be used to provide focus and context  Powerpoint uses this to good effect!

33 33 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Time Visualizations

34 34 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Visual Metaphors : Long time periods  Special techniques have been proposed for visualization over very long time periods  Themeriver technique has been used to depict evolutionary behaviour… .. Possible with Excel Evolution of baby names.... Try it at: http://babynamewizard.com/namevoyager/lnv0105.html Laura and Martin Wattenberg

35 35 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Themeriver  Themeriver for climate change…  …

36 36 18 April 2007 vizNET-LEEDS-PRES-0002-070418 River Metaphor  Taglines  Visualizing tags attached to Flickr online image sharing  Evolution over time  Show tags that are specific to a time period http://research.yahoo.com/taglines/

37 37 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Cluster and Calendar based Visualization of Time Series Data  Jarke van Wijk has shown how visualization can be used in analysis of time series data  Opposite is power demand within ECN (Netherlands Energy Research Centre)…  … hard to pick out patterns of usage

38 38 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Visualizing the Main Clusters  By clustering the 365 days into 6 clusters of similar behaviour, we are able to visualize the key patterns of use…  … but better still, in next slide we link to a calendar

39 39 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Calendar View of Power Demand

40 40 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Calendar View of Number of Employees at Work http://www.win.tue.nl/~vanwijk/clv.pdf What can you observe? (NB Dec 5 th )

41 41 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Spiral Graphs  Spiral graphs are a space- efficient way of visualizing long time series… From Alexa et al

42 42 18 April 2007 vizNET-LEEDS-PRES-0002-070418 Time Wheel  The Time Wheel allows several time series to be viewed simultaneously…  … how successful is this?  … rotation can help, why?  … again cf parallel coordinates? Tominski, Abello, Schumann - Rostock


Download ppt "1 18 April 2007 vizNET-LEEDS-PRES-0002-070418 A Rough Guide to Data Visualization – Part 2 VizNET 2007 Annual Event Ken Brodlie School of Computing University."

Similar presentations


Ads by Google