/* c:\lotteries_b\1997\data\elk_res_1997.dta */

/* The following is the first portion of a program that will comprise 
   the bulk of the do-file demonstration of STATA's data management, 
   statistics and graphical abilities.

   This first portion involves reading a text file into memory, saving 
   it as a STATA data file, generating/transforming some variables, and 
   then examining some numerical and graphical descriptive measures */



/* First, clear the memory, set the space desired, and control the 
   viewing of output */

clear
set memory 80m




/* Second, read in a text (ascii) file and save it as a STATA file */

/*insheet using c:\lotteries_b\1997\data\elk_res_1997_2.txt
save c:\lotteries_b\1997\intermediate\elk_res_1997_2.dta */
use c:\lotteries_b\1997\intermediate\elk_res_1997_2.dta




/* Third, create a "log" file that will store the results */

/*log using c:\lotteries_b\1997\output\stata_presentation.log, replace */




/* Fourth, let's take a quick look at it to see if there are any 
   "problems" and to get an idea of the format of each of the variables */

summarize _all
describe _all




/* Fifth, note that the data is still a bit "dirty" so replace the values
   of some of the existing variables, generate some new variables, and drop 
   some unnecessary variables  */

replace purchase=. if (purchase==-999 | purchase==-888)
replace age=. if (age==-999 | age==-888)
replace age=40.88 if age>97

generate male=1 if gender=="M"
replace male=0 if male==.
/* alternatively: replace male=0 if gender~="M"
                  replace male=0 if gender=="F"          */

generate female=1 if gender=="F"
replace female=0 if female==.

generate obsno=group(_N)

#delimit ;
drop resobsno choice_2 name_l name_f name_m street city state resident gender 
dob phone drawn ;
#delimit cr

/* alternatively: drop choice_2-state resident-drawn */

save c:\lotteries_b\1997\intermediate\elk_res_1997_3.dta, replace
clear




/* Sixth, additional datasets may contain variables of interest for
   the analysis.  If these datasets contain an element that is common
   to our dataset then we can merge the datasets in pairs by the variable
   of interest.

   For example, our dataset contains the zipcode in which each individual
   resides.  Various socioeconomic data is available at the zipcode level,
   so we might want to merge this data with our dataset.

   The objective of the next set of commands is to identify the distance
   an individual will (on average) expect to travel.  To do this, we will
   merge our dataset with an additional data set.  Note that our dataset 
   contains the individuals zipcode and destination choice.  The additional 
   dataset contains 3 variables: origin (individual) zipcodes, destination
   choices, and the round-trip, road distance between the zipcode destination 
   pairs. Since each origin zipcode has 215 destination associated with it,
   the merging below is based upon variable pairs rather than a single, 
   common variable.  */




/* Seventh, we first open the additional dataset containing the mileage 
   estimates, sort it by the two variables of interest, and then save it */

clear
use c:\lotteries_b\shared\elk_miles.dta
rename huntcode choice_1
sort choice_1 orig_zip
save c:\lotteries_b\shared\elk_miles_B.dta, replace
clear




/* Eighth, then we reopen our original, "master" dataset, sort it by 
   the two variables of interest, and then merge it with the sorted 
   dataset from above */

use c:\lotteries_b\1997\intermediate\elk_res_1997_3.dta
sort choice_1 orig_zip
merge choice_1 orig_zip using c:\lotteries_b\shared\elk_miles_B.dta, nokeep




/* Ninth, we then examine the new "master" dataset to make sure 
   nothing "funny" happened */

summarize _all
save c:\lotteries_b\1997\intermediate\elk_res_1997_4.dta, replace




/* Tenth, thus far our descriptive analysis has been at the state-level,
   but we may be interested in examining "stuff" across destination choices. 
   The summarize command is again used but we request that statistics be 
   generated "by" each of the 215 destinations */

keep choice_1 lottery miles age male female
sort lottery
/*by lottery: summarize miles age male*/ 




/* Finally, we conclude this preliminary analysis by examining some 
   graphs.  These are simply intend to complement the summary statistics 
   produced in step 10 */

for num 1/4: erase "c:\temp\gX.gph" 

graph age, histogram bin(10) normal title("Histogram of Age") saving(c:\temp\g1)
graph miles, histogram bin(10) normal title("Histogram of Miles") saving(c:\temp\g2)
graph male female, pie title("Pie Chart of Gender") saving(c:\temp\g3)
graph miles age, twoway title("Scatterplot of Miles vs. Age") saving(c:\temp\g4)

graph using c:\temp\g1 c:\temp\g2 c:\temp\g3 c:\temp\g4