**************************************
* DATA MANIPULATION AND HOUSEKEEPING *
**************************************

set mem 300m
set matsize 200
set logtype text
set more off

log using c:\klaus\REU2009\stata\logs\session_0701, replace
use c:\klaus\REU2009\stata\data\res_monthly


*Preliminaries: Data inspection 
********************************

tab origid
sum use
sum use,detail
tabstat use, stat(mean sd min max)
tabstat use, by(year) stat(mean sd min max)
tabstat use, by(month) stat(mean sd min max)


* Task 1: Create variable labels *
***********************************
label var age "age of building in years"

* YOUR TURN 
* label the variable "bedrooms" as "number of bedrooms" 


*Task 2: Sort data and create a running id 
******************************************
sort origid year month
gen runid=_n

*Task 3: Create a group id for individuals, starting at 1 
*********************************************************
sort runid
egen id=group(origid)

* YOUR TURN 
* create a period id for each month in the time series (1 through 36), call it "period" 


*Task 4: Capture the total number of observations for each individual 
*********************************************************
*(naturally, this will be 36 for all, but in many other applications you may have an unbalanced panel; this is also useful to check if you have a balanced panel, and to quickly eliminate short panels if needed)

sort id
by id: gen panelobs=_N


*YOUR TURN
*do the same for observations per period, call the variable "perobs"


*Task 5: Generate a separate dummy variable for each period
***********************************************************
tab period, gen(p)

* YOUR TURN 
* generate a separate dummy variable for each year. Choose the prefix "y"

tab year, gen (y)

*Task 6: Generate lagged variables
**********************************
xtset id period
gen lag1per=L.period
gen lag2per=L2.period

*Task 7: Re-arrange the order of your variables
***********************************************
order runid id panelobs perobs


*Task 8: Collapse your data to annual consumption levels for each household
********************************************************
* first create household-year id
egen id_year=group(id year)

collapse (mean) id origid year lotsize sqft baths fixtures bedrooms age (sum) use prcp cdd hdd, by(id_year)

save c:\klaus\REU2009\stata\data\res_annual, replace

* YOUR TURN
* collapse further to total use over the three years



*Task 9: Re-load annual data and reshape into wide format (one row per household)
**********************************************************
clear
use c:\klaus\REU2009\stata\data\res_annual

* drop id_year - no longer needed
drop id_year

reshape wide use age prcp cdd hdd, i(id) j(year) 


* YOUR TURN
*back to long format


log close