/* question 1 */
data NO2;
		infile '\NO2.txt' ;
	input logNO2 logV  temp2m vitesse diffTEMP direction heure nbjour;
run;
 
/* question 2 */
 proc sort data=NO2;
 by nbjour;
 run;
 proc print data=NO2;
run;

proc means data=no2 noprint;
var logNO2 logV  temp2m vitesse diffTEMP direction;
by nbjour;
output out=sortie1 mean=mlogNO2 mlogV  mtemp2m mvitesse mdiffTEMP mdirection;
run;

/* question 3 */
data sortie2;
   set sortie1;
   drop _type_ _freq_;
   if mtemp2m>=0 then vit="positif"; else vit="negatif";
proc print data=sortie2;
run;

/* question 4 */
  data positif;
    set sortie2; 
	if vit="positif";   /* 169 observations */

	data negatif;
	set sortie2;
	if vit="negatif";
	run;   /*  118 observations */

	/* question 5 */
	proc freq data=sortie2;
	  tables vit /chisq;
	  run;

	  /* question 6 */
	  proc univariate data=sortie2 normal;
	     var mlogNO2 ;
		 histogram /normal;
		 run;  /* normalité vraie */

		 /*  question 7 */
proc means data=no2 noprint;
    var logNO2 ;
    by nbjour;
     output out=results(KEEP=nbjour medNO2) median=medNO2;
run;

proc print data=results; 
run;

/* question 8 */
data final;
 merge no2 sortie2 results;
 by nbjour;
 dif=medNO2--logNO2;
 run;
 proc print data=final; run;



  /* EXERCICE 2 */
  /* question 1 */
 proc sort data=final;
  by vit;
proc boxplot data=final;
  plot mlogNO2*vit / BOXCONNECT=mean;
  run;

    /* question 2 */
  symbol1 v=plus  c=blue;
symbol2 v=circle c=red;
proc gplot data=final;
	plot ( logNO2 medNO2)*nbjour / overlay legend;
run;
 


/* EXERCICE 3 */

options symbolgen mprint mlogic;
%macro modele(Table,y,z);
proc means data=&table noprint;
    var &y ;
    by &z;
     output out=res  median=medNO2;
run;

  symbol1 v=plus  c=blue;
 proc gplot data=res;
	plot  medNO2*&z ;
	title "graphique de la  variable  MEDNO2 fonction de la variable &z";
run;

%mend;

%modele(NO2, logNO2, nbjour);