/* ECN410 students: use this file to practice using SAS  */

/* These are data described in Chatterjee and Hadi (2012) Regression Analysis by Example.
/* You can find the on Dr. Hadi's website
/* http://www1.aucegypt.edu/faculty/hadi/RABE5/#Download

*/



proc import datafile="g:\Teaching\ECN410\ECN410.DATA\milk.production.xlsx" 
	dbms=xlsx 
	out=work.milk replace; 
	getnames=yes; 
run;

/* get summary statistics */

proc means data=milk;
run;

* check the definitions of the variables at: http://www.public.asu.edu/~rgcox2/ECN410/data/about.milk.production.txt;

* verify that I79 means what the variable definition says;

data milk; set milk;
I79_mine=0; /* create new variable and set it to 0*/
if days gt 79 then I79_mine=1;  /* set my new variable equat to 1 if Days is >79 */
I79_check=I79-I79_mine; /* create another variable as the difference bewteen 
	the original I79 in the data and the variable I created*/
run;

proc means data=milk;
var I79_check;
run;


* What is another way I could have checked this? ;


* If you want to take a better look at the current milk production you can use the proc univariate command. ;

PROC UNIVARIATE data=milk;
var currentmilk;
histogram;
run;

data milk; set milk;
constant=1;  /* create a constant which I will need for the boxplot */
run;


* create boxplots for current milk production;
proc boxplot data=milk;
plot currentmilk*I79;
plot currentmilk*constant;
run;

* test whether there is a statistically significant difference in milk production
between cows that have been lactating for at least 79 days and those that have not;

proc TTEST data=milk;
class I79;
var currentmilk;
run;


* regress the current milk production on the number of days and the number of lactations;

proc reg data=milk;
model currentmilk = days lactation;
output out=a p=predicted r=residual;
run;


proc means data=a mean sum std min max;
var predicted residual;
run;