// *****************************************************************************
// PROGRAM - master OMCA 
// Created by Jane Fry and Clare Boulton, Productivity Commission, Melbourne
// August, 2013.
// *****************************************************************************

// THIS PROGRAM IS A MASTER DO FILE THAT LINKS WITH THREE OTHER DO FILES TO 
// EXTRACT THE HILDA CALENDAR DATA INTO A SINGLE DATA FILE, CONDUCT OPTIMAL 
// MATCHING AND CLUSTER ANALYSIS (OMCA), AND PRODUCE PLOTS AND CHARTS FOR THE 
// RESULTING PATHWAYS.

// NOTE: THIS CODE IS DESIGNED TO WORK WITH HILDA RELEASE 10.0 (CONFIDENTIALISED)
// DATA FILES. THE CODE WILL NEED TO BE ADAPTED TO EXTRACT THE CALENDAR DATA INTO
// A SINGLE DATA FILE FOR OTHER RELEASES OF HILDA DATA. 

// The OMCA1 and OMCA2 programs use commands from the SQ suite written by 
// Brzinsky-Fay, C., Kohler, U. and Luniak, M. 2006, 'Sequence analysis with Stata',
// The Stata Journal, vol. 6, no. 4, pp. 435-60.  
// Running the following command will download the most recent version of this Stata ado-package:
// ssc install sq

// Alternatively, the SQ program suite can be downloaded from the following website 
// (maintained by Brendan Halpin):
// http://teaching.sociology.ul.ie/seqanal/


// Before running the programs users must specify directories and the file 
// pathway for the HILDA data. 

// specify location of datafiles, programs and results
cd "H:\Hilda R10\100c"

log using "Master_OMCA.log", replace

// The clustering commands below and subsequent descriptive analysis is set up for 
// youths with 5 pathways. For a different age group, or with a different number 
// of pathways, changes are needed as indicated below and in the other programs 
// (OMCA1 and OMCA2).

// *****************************************************************************
// Aggregate HILDA data into monthly data, then merge into a single wide dataset
// *****************************************************************************

do "10yrs HILDA calendar" // This program creates the file "activity_month_wide.dta"
						  // containing the calendar data for analysis.

// *****************************************************************************
// Optimal matching and cluster analysis example: YOUTHS
// *****************************************************************************

use "activity_month_wide.dta", clear

// Restrict sample to youths (for example)
keep if w1age==1

// ssc install sq // run this command to install SQ program suite

// Perform optimal matching and generate a dendrogram
do "OMCA1"	

// Save and analyse dendrogram for youths
graph save "youth dendrogram", replace

// Decide on the number of clusters and generate indicator
// variable for the pathway number (5 pathways in this example)
cluster generate groupWARD_5=groups(5), name(distWARD) 
sqclusterdat, return 

// Create pathway labels to use
label define pathno 1 "pathway 1" 2 "pathway 2" 3 "pathway 3" ///
4 "pathway 4" 5 "pathway 5" 

label values groupWARD_5 pathno 

// Descriptive analysis (sequence index plots, determine the concentration of 
// sequences in the data & produce data for chronographs)
do "OMCA2"

log close