What about Swivel

1
What about Swivel?

• Amy West
• ICPSR Official Representatives Conference
• October 6, 2009

2
What is Swivel?
3
Typical Situation

• Consumer Expenditure (CEX) Survey
• Very small part of regular management and
  marketing class assignments
• Native presentation sub-optimal for comparing CEX
  columns, e.g. expenditure patterns for different
  age groups for different products

4
CEX Table Output

• Each column is separated into a different table
• To create a single table with all selected
  columns will mean time spent with a spreadsheet
  program by librarian/instructor/students
• Effort far exceeds points in assignment
• Can Swivel bypass this part of the process?

5
Swivel Upload Requirements

• By data, they mean table
• All relevant information must be in single table,
  with representation on each row or column
• Additional content, like footnotes, have no space
  in Swivel

6
Foreshadowing

• Data files should contain data
• http//blog.plover.com/misc/census-data.html

7
CEX Flat Files Look Like This

• CEX files are relational, e.g. not all relevant
  information is in one table
• Footnote codes are stored as numbers, but they
  arent data and they are in the file
• Wackiness in the series_id makes cleanup on this
  file more hassle than it should be.

8
What if the data were ready/easy?

• Cons
• Layout, labeling
• Retaining changes to graphs
• Export of graphs
• No API (yet)

9
Graph Example

• Title becomes Qtr3
• Metadata separated from graph image

10
However
11
So, what about Swivel?

• As is, extremely limited application in an
  academic context
• Too large a gap between what Swivel accepts and
  what is available
• Swivel positions itself as a destination
  academic institutions need tools that can be
  integrated into their environment
• Metadata and export problems with graphs

12
And ManyEyes, Google Spreadsheets, etc?

• Only Google Spreadsheets seems to have much of an
  advantage and its a slim one.
• Google has better data intake from live sources,
  but charts, et al are still constrained by the
  data layout.
• Neither Google Spreadsheets or ManyEyes includes
  citation metadata in browser.

13
What might make a difference?

• Swivel API that integrates with course software
  like Moodle
• Data files that contain only data
• Semantic markup of data

14
Conclusion

• Casual to mid-level data users still need a
  middle-ware solution to the high-barrier data
  available
• Swivel et al arent that solution in that they
  dont enter the workflow until after most of the
  hard work would have to be done.

15
Thank You

• Amy West
• Data Services Librarian
• [phone]
• westx045_at_umn.edu
• http//www.lib.umn.edu/data
• Google Mail, IM, Docs, Reader ae.west
• http//delicious.com/umdatalib
• http//twitter.com/amyewest