Kelp Diagrams: Point Set Membership Visualization Kasper Dinkla,Marc J.van Kreveld

1. Kelp Diagrams: Point Set Membership VisualizationKasper Dinkla,Marc J.van Kreveld Qinglai He Vision Computing Lab,SCU 2013/11/01

2. 1 2 3 4 5 Index Introduction Problem Analysis Approach Results and Discussion Conclusion 2/23

3. 1. Introduction • Complex • Invalid set membership Bubble Sets &LineSets • Allocation of space around each element • Allocation of space for connecting shapes b • Genetationg of actual visualizations Kelp Diagram 3/23

4. 1. Introduction • Two styles of diagram that emphasize different aspects of set memberships and overlap for elements with a predefined position; • A routing algorithm for linking elements in a set to support the generation of such diagrams, where aesthetic quality, efficiency, and effectiveness are taken into account. Contribution 4/23

5. 1 2 3 4 5 Index Introduction Problem Analysis Approach Results and Discussion Conclusion 5/23

6. 2. Problem Analysis • Input problem instance • positoned elements • the sets that contain them • the predefined visualization that embeds the elements • Task drived • Supported tasks • Constraints • Aesthetic criteria 6/23

7. 2. Problem Analysis Supported tasks T1a determine the position of an element T1b find an element by position(relative to landmarks) T1c estimate the density of elements in an area T2a determine which sets contain a specific element T2b find the elements that belong to a specific set T2c estimate the spatial distribution of a specific set T3 compose a (mental) set from existing sets with operations union, intersect and complement ,and apply T1-T2 7/23

8. 2. Problem Analysis Constraints C1 every element is clearly represented in the final visualization at its predefined position ( for T1) C2 every element is clearly marked or contained by a representation of every set that it is a part of ( for T2) 8/23

9. 2. Problem Analysis Aesthetic criteria A1a small area A1b few and shallow bends A1c few outline intersections A2a large area A2b large distance between outlines A2c little overlap of shapes that depict different sets A2d strong continuation of shapes that depict the same set A3a little obfuscation of the predefined visualization A3b strong correspendence between the presence and size of a set's shape , and the presence and density of elements that belong to this set, in an area 9/23

10. 1 2 3 4 5 Index Introduction Problem Analysis Approach Results and Discussion Conclusion 10/23

11. 3. Approach • Three phases • allocation of element space -----> Element space ( phase 1) • allocation of additional link space -----> Link space ( phase 2) • the generation of visualizations -----> Visual styles ( phase 3) 11/23

12. 3. Approach phases 1. Elment space 12/23

13. 3. Approach phases 2. Link space 13/23

14. 3. Approach phases 2. Link space 15/23

15. 3. Approach phases 2. Link space 14/23

16. 3. Approach phases 3. Visual styles ( Nested style & Striped style) 16/23

17. 1 2 3 4 5 Index Introduction Problem Analysis Approach Results and Discussion Conclusion 17/23

18. 4. Results and Discussion 18/23

19. 4. Results and Discussion 19/23

20. 4. Results and Discussion 20/23

21. 1 2 3 4 5 Index Introduction Problem Analysis Approach Results and Discussion Conclusion 21/23

22. 5. Conclusion • Contribution • Balances visual complexity,based on aesthetic criteria,with effective depiction of the data • Kelp Diagrams hava a consistent, easy to interpret, appearance • Flexibility that is required for application to different kind of visualization • Future work • The routing algorithm is too slow for interactive use • Supporting elements with dimension ( instead of points) and automated derivation of parameter settings based on the data itself. • Explore a more generic method 22/23

23. Thank you for your attention!