EDA: Exploratory data analysis
1. Read the Protocol
2. Familarize the CRF
3. Start EDA with the data file 
4. Clear goal in mind=> the primary outcome (DV) and, probably, the IDV of interest

Stata commands:

	clear
	set more off
	cd "C:\DMHT_2553_54_55\"
	use "07_DMHT2554_Data_STATA.dta" , clear

   * Continuous outcome -> dichotomize with a complex definition
	desc
	lookfor 18
	li b1811 b1812 in 1/20
	li b1811 b1812 b1821 b1822  in 1/20
	replace b1811 = . if b1811 < 30
	replace b1812 = . if b1812 < 30
	replace b1821 = . if b1821 < 30
	replace b1822 = . if b1822 < 30
	gen bpsys = (b1811 + b1821)/2
	gen bpdia = (b1812 + b1822)/2
	li b1811 b1812 b1821 b1822 bpsys bpdia  in 1/20

	gen bpyn = .
	replace bpyn = 1 if bpsys <130 & bpdia < 80 /*Need Ref*/
	replace bpyn = 0 if bpyn !=1 & (bpsys !=. | bpdia != .)
	li b1811 b1812 b1821 b1822 bpsys bpdia bpyn  in 1/20
	tab bpyn

	tab opd
	tab bpyn
	tab opd bpyn
	tab opd bpyn, row
	xi: logistic bpyn i.opd

* Continuos outcome longitudinally
	use "09_DMHT2553_54_55_Data_Final_20130522.dta", clear
	tabstat dm_hba1c, stat(n mean sd median min max) by(year)
	tabstat dm_hba1c, stat(n mean sd median min max) by(type)
	tabstat dm_hba1c, stat(n mean sd median min max) by(sex)
	ttest dm_hba1c, by(sex)
	regress dm_hba1c sex