1. UFinder An interactive tool for data exploration and decision-making Jeanette Bautista Micheline Manske

4. Problem There are many variables to consider when choosing a school Everyone assigns weights these variables differently Solution Visualization that allows users to select ranges for variables to reduce the number of schools Interactive tool to allow users to weight variables differently and explore how this changes the rankings of schools

5. Data set 1300 US colleges and universities 33 variables Variables are arranged hierarchically

6. Caliber of students marks mathverbalACT Cost / year schoolliving bookstuitionadditional fees % top 25graduation rate room & board room spending Environment student / faculty ratio total # students full time part time Acceptance % new students enrolled / accepted % applications received / accepted

7. Initial Design

9. Problems encountered Some data values missing Slider layout may use up too much screen space

10. Caliber of students 0 - 100 Cost / year 0 - 100 000 Environment 0 - 100 Acceptance 0 - 100 0 50 000 100 000

11. Caliber of students 0 - 100 Cost / year 0 – 100 000 School 0 – 60 000 Living 0 – 40 000 books 0 - 5000 tuition 0 - 50 000 additional fees 0 - 5000 room & board 0 – 25 000 board 0 – 10 000 room 0 – 15 000 personal expenses 0 – 15 000 Environment 0 - 100 Acceptance 0 - 100 0 25 000 50 000

12. Caliber of students 0 - 100 Cost / year 0 – 70 000 School 0 – 30 000 Living 0 – 40 000 books 0 - 5000 tuition 0 - 20 000 additional fees 0 - 5000 room & board 0 – 25 000 board 0 – 10 000 room 0 – 15 000 personal expenses 0 – 15 000 Environment 0 - 100 Acceptance 0 - 100 0 25 000 50 000

13. Problems encountered Some data values missing Slider layout may use up too much screen space Combining values hierarchically may be difficult

14. Caliber of students marks mathverbalACT Cost / year schoolliving bookstuitionadditional fees % top 25graduation rate room & board room spending Environment student / faculty ratio total # students full time part time Acceptance % new students enrolled / accepted % applications received / accepted

15. Problems encountered Some data values missing Slider layout may use up too much screen space Combining values hierarchically may be difficult it depends on how much you value each component

16. Demo

17. Problems with ValueCharts A value for a variable must represent a positive aspect in ValueCharts. - inverse relation There are too many schools to list them all - sliders to narrow range There are too many variables to have a column for each - multiple value charts

20. Suggestions?

21. Where we’re at Become familiar with the InfoVis toolkit Discuss ideas with Giuseppe Carenini Implement preliminary interface (no functionality)

22. Where we’re going Get sliders working with a scatterplot Implementation with ValueCharts Incorporation of multiple levels of ValueCharts Extension?

23. References [1] Giuseppe Carenini and John Lloyd. ValueCharts: Analyzing Linear Models Expressing Preferences and Evaluation. In publication. [2] Jean-Daniel Fekete. The InfoVis Toolkit. Version 0.6alpha2, 2004. http://www.lri.fr/~fekete/InfovisToolkit/ [3] US News “colleges” data set. 1995 Data Analysis Exposition sponsored by the Statistical Graphics Section of the American Statistical Association. http://lib.stat.cmu.edu/datasets/ http://lib.stat.cmu.edu/datasets/