Notes on running the TA Matching program (Longer Explanation)

create the input file ("preferences.txt")

preferences.txt is the input file for the hospital/residents (i.e., courses/ta's)
matching program.

to create this file, use the portal webapp, TAM: Generate TA Prefs Input File. It
will take about 5 minutes for the screen to change to the download
screen where you can download the file. You may have to use Firefox
for this (Internet Explorer may "time out") since it takes so long
waiting for the file to be created.

download and save the preferences.txt file to your local computer's
hospital_residents folder.

FORMAT: the preferences.txt input file that you create consists of two
types of lines: - "C:" lines (courses/profs, and their TA prefs), and
- "T:" lines (TA's and their course prefs).

the format of each is:

C:title,instructor,course_num(socsci),num_tas_applied,course_code(registrar),dept,num_ta_slots,ta_pref1,ta_pref2,ta_pref3,...

(the ta's are specified by their "ss_id" -- an ID number we just
use for them here in social sciences computing services for our
database).

T:last_name,first_name,ss_id,student_id,ucinetid,commitment,course_pref1,course_pref2,course_pref3,...

knowing this about the file, i scan it, taking a sample of a few of
the lines, just to check things out. the kinds of things i look for
are:

- are these the courses you know to be offered in that qtr?

- do the right number of ta slots show up for each course?

- is the right prof listed for each course?

- did the prof's preferences get registered (do they show up on the
C: line)?

- did the TAs who were excluded¹ for this qtr truly get excluded
(i.e., they do NOT show up in any T: line)?

- are commitments (true,false) correct?

- are there any TAs listed that you now realize you should go back
and exclude¹?

- did the "pre-matched"¹ (forced match) TA's go into the courses they
were pre-matched to?

- are there any TAs who listed very few course prefs (their T: lines
are extremely short)? these students probably put in a very
restrictive schedule.

- are the student id's correct (a la our latest issue with duplicate id's)?

and so on..

run the matching program to create the stablematch.txt output file


to create the output file containing the matches of students to
courses, double-click on the run.bat program from your
hospital_residents folder on your computer. this will run the Java
program which is contained inside the "hospital_residents.jar"
file.

when the hospital/residents program window comes up, click on the
"Complete" button. Wait a few minutes until it is done. When it is
done, the button will pop back up, and you can close the window. Your
new stablematch.txt output file should now be created (or updated) in
the same folder.

view the stablematch.txt file for its separate reports.

if you want, copy/paste the first (CSV) section to a separate text
file and then import that text file into an excel spreadsheet.

  - after the CSV section there are 4 separate reports in the
    stablematch.txt output file: "prefsA", "prefsB", "prefsC", and
    "prefsC". these were all created for Diane E. on previous
    quarters, per her request. some of these reports get quite long,
    so you might want to search forward for them in your editor, using
    Ctrl-F, searching for "prefs". this will quickly take you to the
    next report.

Re-run the process


if you decide to exclude a student who has been given funding
elsewhere, or remove from exclusion a student, or to pre-match (force)
a student into a course, you can use the portal webapp, TAM: Tune TA Match Processing to do so.

once you've done this re-tuning, you can run steps 1 and 2 again.

it is better to make these manual changes BEFORE steps 1 and 2, if at
all possible. making manual changes AFTER steps 1 and 2 will result in
a less "smooth" fit, and less happy people. we want happy people har
har ;)

==================
¹using the portal webapp, TAM: Tune TA Match Processing

Back