OFIS Notes from 15 May 2009

From April 29 meeting:

• Daniel presented about CS course critiques.
  • data is passed to Frank Tompa from...? -MFCF, from Math Undergrad Office and IST
  • Engineering: data goes from IST -> Karen Dubois [Dean of Engineering Office] -> excel -> SAS for analysis -> Marlon in .csv format
  • Questions on CS data: (Daniel will ask Frank, as necessary)
    • Teaching roster comes from where? -see below
    • Do we get a person ID anywhere? -No
    • cross-listed courses: should they be listed as separate rows? -Not currently done that way
    • lab section: / : show, or not? think we don't care. -Don't care currently
      • Preferred data formats: <08.3> -> semi-julian date: 1089 (months: 1,5,9, 7) -Up to us
    • CS 100 1 -> CS, 100, 1 -Yes, feasible

This week, I talked to Lori Suess about student evaluations.

She (actually Jim Johnston) receives two data-sources each term: excel from Math Undergrad Office; and a floppy disk of data from IST. She sent me a sample of each. These are merged with a canned output-template to produce the evaluation pages, in .ps, text, and ([t]roff) format. That output is sent to Frank Tompa (and the student math office; for the book). It is not archived by Math.

The floppy from ist contains a text file with individual coded student responses

517000001001041907001   5363 #0001    N 51001                
517000002001041907001   5363 #0001    N      232316211223332 
517000003001041907001   5363 #0001    N      132216221222321 
517000004001041907001   5363 #0001    N      131 26211132331 
517000005001041907001   5363 #0001    N      231116311223322 
517000006001041907001   5363 #0001    N      232116211222332 
517000007001041907001   5363 #0001    N      122116221223313 
517000008001041907001   5363 #0001    N      232216211232211 

[...]

517000083001041907001   5363 #0001    N      231233 22222323 
517000084001041907001   5363 #0001    N      231236221222331 
517000085001041907001   5363 #0001    N 51002                
517000086001041907001   5363 #0001    N      523446253222232 

[...]

First row denotes a class; following rows are individual student responses, up until the non-indented delimiter "51002".

• delimiter-line: [card number from student] [garbage data; always the same] [course number 51001]
  • course number 51001: 5 means it's cs.
• data lines: course-number is replaced by student answers to each question in order.

Excel data from Math Undergraduate Office

Index #	Subject	Catalog	Sect	Last Name	First Name	Middle	Tot Enrl	Dept.
12311	ACTSC	231	001	Chisholm	Diana	Katherine	157	MC 6136
12321	ACTSC	232	001	Chisholm	Diana	Katherine	92	MC 6136
13311	ACTSC	331	001	Cai	Jun		105	MC 6136
13711	ACTSC	371	001	Kim	Hyun Tae		126	MC 6136
13721	ACTSC	372	001	Hardy	Mary	R	155	MC 6136
14331	ACTSC433/ACTSC833		001	Li	Siu Hang		61	MC 6136

[...]

• index# corresponds to the 51001 in the other file.

How does this turn into the free-form text?

• Lori runs a checksum program, she believes written by wcwince; it has been updated by Jim J any time Math Undergrad Office changes their data-stream; but they aren't very fond of updating it. Ray has written a script in .csh to try and do the same thing; but it isn't in production.
• David Mcinnon, Math chair for undergrad studies, also was working on excel data-merging tool, but it was last year and he hasn't gotten back to them about getting the real data. The perl script she emailed me appears to be similar to capo:/fsys2/source/course-evaluation/ceqsum/ceqsum

bash-3.00$ ls -la /fsys2/source/course-evaluation/ceqsum/ceqsum ceqsum-test.pl
-r-xr-xr-x   1 root     none       23004 Apr 13  2005 /fsys2/source/course-evaluation/ceqsum/ceqsum
-rw-r--r--   1 drallen  other      24923 May 14 13:43 ceqsum-test.pl
bash-3.00$ diff -w /fsys2/source/course-evaluation/ceqsum/ceqsum ceqsum-test.pl |wc  
     126     525    3819

The diff appears to include input-data differences as well as output-formatting. As listed at the header of both files, capo:/fsys2/source/course-evaluation/ceqsum/ceqsum.perl is (lightly) commented.

Lori suggested that she would be happy to email us a copy of the .txt and excel data each term. They would be pleased to work with me/us to find some solution that feels less clunky than the current system.

Past Tasks:

• response from MFCF about processed data. [above]

• Figure out which Engineering prof(s) and admin(s) to talk to ; maybe do a few SQL queries on DB to see who is using it?
  • looked at "Academic_background" as a sample SQL table to look for recent users; mapped Person_id to Faculty_member. Found my own entry; three subsequent entries. All three are outside engineering: these people are in Philosophy, English Language, and Recreation and Leisure Studies. In the set of users just before mine, two are CS prof and BFA. And two are engineering profs. Emailed them.
  • Possibly ask Marlon who in the Dean's office would be good to ask?
    • Looked at Person_delegate table; a few recent additions were delegated on behalf of lots of chemical engineering profs. Did not contact them.
  • waiting for any response from the two Engineering profs I contacted.

Upcoming Tasks

• Daniel: Further investigation of http://www.tug.org/utilities/texconv/textopc.html

-- DanielAllen - 14 May 2009