quantum sensing trends project andrew wharton december 8, 2012
TRANSCRIPT
Quantum Sensing Trends Project
Andrew WhartonDecember 8, 2012
Objective
• Visualization for the USAF to detect trends for Quantum Sensing research
• Easy, efficient tool• Provide more than one perspective
Data
• One excel spreadsheet– Poor labeling– Formatting issues
• One BibTex file– Tough to export into CSV format• Used Mendeley to export to RIS• Then used JabRef to convert RIS into CSV
• *CSV is format desired for d3 use
Parsing and Filtering
• Inconsistent format between two files• Also too much data—needed to filter• Imported CSV into MySQL DB– Used queries to filter data– Export query results back to CSV
• Import CSV into d3
Initial Design
• Animated bubble chart using d3• Bubble represents keyword, size indicates
frequency
Problems
• Too many years to share screen space– Unable to distinguish bubbles
• Could not easily detect trends over time– Identified this as a potential issue in the beginning,
but decided to use colors to group bubbles across years
– But, there were too many keywords, so colors were not unique
• Programming bugs– Animation was not working
Re-Design
• Keep it simple, but try something new using existing systems
• Create multiple components using Tableau, ManyEyes
• Integrate these components to provide user with opportunity to explore and discover in multiple ways
Idea
• Four components– Tag Cloud
• Provides overview
– Bubble Chart• Also provides overview, but in a different way
– Bar Graph• Shows breakdown over the years
– Line Graph• Easier way to identify trend over time
• All sortable by year, keyword(s)
Result
• Not quite the “wow” factor as initial design• Still answers questions • Decided to discard BibTex data– Too few results, skewed the data and portrayed an
inaccurate visualization• Lessons learned with design and development– New frameworks = LOTS of overhead
• If few existing examples, re-prioritize based on time
– Don’t go against basic design principles• Being new is great, but a backup plan is needed in event of failure