Visualizing the Intersections of Data

Vaughn Sousa

I was inspired by some of the visualizations that our research faculty produce and wanted to see if I could reproduce it in Oracle Analytics. See the link I posted for an explanation of the use of an Upset Chart since it is better than I could put to words here.

1. Which dataset did you use?
   • An anonymized residency-application dataset containing one row per student and a set of boolean columns (Y/N) for each specialty (e.g., Internal Medicine, Neurology, Pediatrics), plus a student key and a few helper flags. This structure supports computing overlaps across specialties for an UpSet view of co-application patterns.
2. How did you analyze or prepare the data?
   • Shaped for UpSet: kept one row per student; ensured each specialty column is strict boolean (Y/N → 1/0); filtered rows with incomplete specialty responses.
   • Created a simple measure for counting students and an optional measure for weighting (e.g., GPA band or priority) to demonstrate Sum vs. Count aggregation.
   • Built a custom UpSet Chart Oracle Analytics plugin to render intersections and totals. Key enhancements:
     • Redwood theme palette and named palette support; rounded corners and increased bar spacing.
     • Perfectly centered initials inside bubbles; counts displayed above intersection bars; per-mark tooltips (counts and percentages).
     • Grammar options: tooltip toggle, color override, legend menu (using delivered legends); field legend hidden in the visual to reduce clutter.
     • Responsive layout to keep bars and labels within the canvas; improved hover behavior so tooltips don’t linger.
3. Who is the intended audience for your visualization?
   • Medical school leadership (deans, program directors) planning interview capacity and outreach.
   • Specialty departments assessing co-application patterns and competitive overlaps.
   • Student advising teams helping candidates build balanced specialty portfolios.
   • Research Data analysis
4. What is your visualization about, and what question or problem does it address?
   • The UpSet chart reveals which specialties students commonly apply to together, the size of each intersection, and how those intersections compare to individual specialty totals. It addresses scheduling, resource planning, and advising questions such as:
     • Which specialty pairings or trios are most common?
     • Where are potential interview conflicts likely to happen?
     • Are certain cross-application patterns increasing year over year?
5. Did you use any Oracle Analytics AI features when building your visualization (e.g., AI Assistant)? If so, please describe how they were used.
   • No. Data prep used standard OA expressions/filters, and the visualization is a custom UpSet plugin developed for Oracle Analytics.

Benjamin Arnulf-Oracle

Love it! Thank you for sharing with the community Vaughn!