/* Data and program from 
   "Analyzing activity tracker data: How many steps per day do YOU take?"
   by Rick Wicklin
   http://blogs.sas.com/content/iml/2014/08/22/activity-trackers/
*/

/* Data from tweets during a 10-minute period 
   for activitiy recorded on on 8/17/2014 
   for a particular fitness activity tracker */
data Tracker;
informat Steps COMMA. Miles;
format Steps comma.;
input Steps Miles;
datalines;
4,314  1.8 
4,635  1.9 
11,343  4.6 
4,991  2.1 
8,708  2.3 
12,469  5.1 
11,863  5.5 
26,686  12.7 
8,283  3.5 
3,533  1.5 
4,580  1.8 
14,029  6.4 
21,606  12.8 
18,069  13.5 
1,104  0.5 
4,348  1.9 
4,804  2.3 
9,452  4.1 
13,672  6.1 
7,639  3.5 
5,091  2.3 
3,482  1.6 
12,731  6.3 
9,729  4.4 
6,209  2.8 
5,508  3.1 
28,469  13.4 
8,220  3.5 
7,922  3.7 
7,728  3.3 
9,084  3.8 
4,038  1.8 
6,653  2.7 
7,999  3.3 
2,925  1.3 
1,996  0.8 
7,475  3.6 
15,584  6.9 
16,246  12.2 
5,872  2.6 
5,159  2.4 
10,844  5.3 
13,116  6 
12,273  5.5 
7,460  3.1 
5,594  2.3 
3,226  1.4 
5,764  2.5 
953  0.4 
7,000  3.1 
13,486  6.2 
2,979  1.3 
3,070  1.2 
7,371  3.5 
14,090  6.6 
420  0.2 
11,176  2.8 
7,532  3.1 
6,889  2.8 
1,958  0.9 
11,520  6.5 
4,821  2.3 
10,755  4.1 
4,003  1.7 
6,175  2.9 
53  0 
2,983  1.2 
17,375  12.4 
13,573  6.5 
2,539  1.1 
13,240  5.5 
12,159  4.4 
2,216  1.1 
13,009  2.3 
12,216  5.4 
3,630  1.6 
7,191  3.3 
13,555  5.8 
7,705  3.3 
73  0 
8,526  3.6 
6,207  2.8 
12,274  5.4 
7,319  3.3 
10,424  4.3 
14,248  6.2 
1,794  0.8 
10,975  4.7 
3,774  1.6 
11,822  4.9 
6,452  2.8 
8,472  3.9 
1,013  0.4 
2,501  1.1 
12,643  5.3 
3,279  1.4 
5,252  2.2 
1,541  0.7 
4,472  1.7 
8,425  3.6 
10,027  4.4 
1,784  0.7 
2,929  1.5 
7,848  3.3 
2,612  1 
3,642  1.7 
3,686  1.6 
16,102  7.6 
2,703  1.2 
2,107  0.9 
9,634  4.4 
5,022  2.2 
14,209  6.3 
5,922  2.4 
4,341  1.9 
6,415  3.1 
372  0.2 
10,378  4.7 
9,131  4.2 
4,518  2 
6,685  3.3 
7,978  3.3 
8,893  4.1 
7,090  3.1 
6,731  3.1 
12,507  5.5 
3,648  1.4 
8,659  3.4 
15,251  6.8 
11,349  5.2 
10,381  4.4 
16,675  7.2 
3,101  1.4 
4,594  1.9 
11,006  6.1 
5,028  2.4 
18,019  8.8 
3,661  1.6 
23,684  20.2 
4,218  2.4 
10,499  5 
11,856  5.2 
9,551  4.2 
6,422  2.7 
1,705  0.7 
5,744  2.2 
2,492  1 
13,906  6.3 
5,628  2.6 
6,557  2.9 
3,518  1.5 
17,107  7.6 
2,546  0.7 
4,203  1.8 
18,468  8.7 
11,601  4.9 
4,664  2.1 
15,841  7.7 
99  0 
16,526  7.8 
9,587  5 
503  0.2 
3,394  1.7 
3,369  1.4 
4,012  1.5 
5,361  2.4 
12,771  5.4 
7,107  3.7 
3,152  1.3 
9,629  4.5 
3,291  1.3 
7,997  3.6 
3,168  1.4 
17,202  8.1 
5,161  2.3 
2,420  1 
18,831  8.6 
5,245  2.3 
10,757  4.3 
4,865  1.9 
20,676  9.5 
6,066  2.8 
5,055  2 
5,674  3 
1,652  0.6 
12,904  6 
3,450  1.4 
2,118  0.9 
8,233  3.5 
4,792  2.2 
6,820  4.8 
1,880  0.8 
41,734  16.9 
11,054  4.8 
5,913  2.4 
5,709  2.5 
13,910  6.5 
3,156  1.4 
27,828  13.1 
14,920  6 
2,264  1.1 
8,320  3.7 
4,519  1.9 
12,552  5.7 
10,288  4.4 
10,038  4.3 
1,977  0.9 
4,275  2 
10,278  4.6 
1,139  0.5 
2,365  1 
7,106  3.4 
8,953  5.1 
1,421  0.4 
5,346  2 
1,874  0.8 
11,858  4.8 
14,413  6.7 
12,459  8.8 
8,068  3.6 
13,915  10.9 
13,398  6 
26,255  14.1 
10,382  4.8 
11,067  5 
11,491  5 
11,507  5.4 
5,376  2
14,636  5.6 
20,389  8.9 
6,850  2.9 
;

proc means data=Tracker N mean Q1 median Q3 QRange;
var Miles Steps;
run;

title "Distribution of Steps for 231 Users";
proc univariate data=Tracker;
var Steps;
histogram Steps / endpoints kernel(lower=0)
                  odstitle=title odstitle2=title2;
inset N Q1 median Q3 P95 / pos=NE;
run;

title "Miles vs. Steps from Tracker Data (N=231)";
proc sgplot data=Tracker;
where Steps<=20000;
scatter x=Steps y=Miles;
xaxis grid; yaxis grid;
run;

/* fit robust regression lin. Slope is average "Miles per Step" */
proc robustreg data=Tracker method=lts fwls outest=RROut;
model Miles = Steps / noint;
run;

data Results;
set RROut;
MilesPerStep = Steps;
StepsPerMile = 1/MilesPerStep;
FeetPerStep = 5280 * MilesPerStep;
CmPerStep = 30.48 * FeetPerStep;
keep StepsPerMile FeetPerStep CmPerStep;
run;

proc print data=Results noobs; 
run;