• DataProcessingAndNumericalAnalysis

Session Title: Data Analysis and Numeric Processing Organizer: Patrick Ball

(someone PLEASE format this later...)

Attendees:

1. Phil Hodge
2. JC Hsu
3. Perry Greenfield
4. R Kaplan

5. M McFarland

6. Rick Grubin
7. Robert Jedrzejen
8. Jeff Mathews
9. David Brown
10. Erin Stantill
11. Abhay Saxena

Numeric doesn't do what we want (although it's great for what it does do)

Patrick likes Stata

• Many things can do what we want in SQL

A story by Patrick Multiple systems estimation

• Find intersections of lists
• Create list of presence/absence in the other lists
• Estimate what lies outside

Suggestion from Numarray folks

• Use recarray
• Mutable values, not mutable shape
• Perry describes record arrays in detail (sorry, I didn't get this down...)

Suggestion by Jeff/Abhay

• Use Metakit http://www.equi4.com/metakit/

Use R (like S-Plus)

• http://www.r-project.org/

• http://cran.r-project.org/

• http://rpy.sourceforge.net/

• http://www.omegahat.org/RSPython/index.html

R + Python has problems. Patrick tells of difficulties passing data between Python and R, even simple stuff like small matrices -- R gives back data errors (roughly: "I don't understand what this is").