UFinder An interactive tool for data exploration and decision-making

1. UFinderAn 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 Cost / year school living marks % top 25 graduation rate books tuition additional fees room & board math verbal ACT spending room board Environment student / faculty ratio total # students Acceptance full time % applications received / accepted % new students enrolled / accepted part time

7. Initial Design

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

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

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

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

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 Cost / year school living marks % top 25 graduation rate books tuition additional fees room & board math verbal ACT spending room board Environment student / faculty ratio total # students Acceptance full time % applications received / accepted % new students enrolled / accepted part time

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/