Stat 200 Final Project

How does exercise intensity affect sleep quality?

By Leland Krych

First we need to read in the data!

Activity<-read.table('/Users/Leland/Desktop/Fitbit Data/Stats Final Project.csv',header=T, sep=",")
head(Activity)
##    X            Date Activity X.1 Distance..Mi. Duration..Min.     Rate
## 1 NA   Nov 30 8:07AM      Run  NA          3.34          22.13 6.625749
## 2 NA Nov 28 11:01 AM      Run  NA          3.39          22.65 6.681416
## 3 NA  Nov 24, 8:16AM      Run  NA          3.97          22.35 5.629723
## 4 NA  Nov 22, 8:57AM      Run  NA          3.48          21.90 6.293103
## 5 NA  Nov 20, 4:49PM      Run  NA          4.44          28.73 6.470721
## 6 NA  Nov 18, 8:18AM      Run  NA          4.18          23.42 5.602871
##   Time.of.Day Minutes.Asleep Minutes.Awake Ratio.of.MinAwake.MinAsleep
## 1          AM            440            28                   0.9401709
## 2          AM            518            34                   0.9384058
## 3          AM            460            24                   0.9504132
## 4          AM            452            26                   0.9456067
## 5          PM            395            26                   0.9382423
## 6          AM            432            34                   0.9270386
##   Weekday.
## 1        Y
## 2        N
## 3        Y
## 4        N
## 5        Y
## 6        Y

We have a total of 46 instances of runs.

Now that we’ve read in the data let’s take a look at the variables that are going to inform our analyses. We want to look at the distance I ran and the duration of the runs because we will use the two of these two make a variable that is equal to the average pace of my run.

distance<-Activity$Distance..Mi.
hist(distance)

Notice that more than half of the runs were in the 5k distance.

duration<-Activity$Duration..Min.
hist(duration, xlab = "Time spent running")

Again, we see that more than half of the runs are in between 20 and 30 minutes.

Cool, let’s investigate the pace of the run.

RunPace<- duration/distance
mean(RunPace)
## [1] 6.865468
sd(RunPace)
## [1] 0.6754506
hist(RunPace, main = 'Speed of run(min/mile) between Aug-Nov 2015', ylab = 'Frequency', xlab = 'Speed of run (min/mile)')

Great, let’s look at my sleep statistics now.

hours_asleep<-Activity$Minutes.Asleep/60
hist(hours_asleep)

Divide by 60 so that results are stored in hours not minutes.

hours_awake<-Activity$Minutes.Awake/60
hist(hours_awake)

QualityofSleep <-hours_asleep/(hours_asleep+hours_awake)
hist(QualityofSleep, main = 'Ratio of time asleep vs total time in bed between Aug-Nov 2015', ylab = 'Frequency', xlab = 'Percent time asleep')

Let’s plot the two against eachother.

plot(QualityofSleep,RunPace, main = 'Comparison between quality of sleep and run intensity', ylab = 'Run intensity (min/mile)',xlab = 'Percent time asleep')
abline(lm(RunPace~QualityofSleep), col="red")

cor(QualityofSleep,RunPace)
## [1] -0.2613655

Pace and Sleep quality are negatively correlated, something I would have expected, but the relationship isn’t very strong.

There are many things that could have influenced this data: Caffeine/Alcohol intake, The time of day that I ran, the amount of stress I was feeling at the time, whether the run was on a weekend vs a weekday.

I investigated a couple of these confounders and further subdivided the data by them.

Now let’s divide the runs by time of day and weekend vs weekday.

ColumnsYouNeed<-c(Activity$Rate,Activity$Ratio.of.MinAwake.MinAsleep)
Morning<-Activity[Activity$Time.of.Day == "AM", c("Rate","Ratio.of.MinAwake.MinAsleep","Distance..Mi.")]
Afternoon<-Activity[Activity$Time.of.Day == "PM", c("Rate","Ratio.of.MinAwake.MinAsleep","Distance..Mi.")]
head(Morning)
##       Rate Ratio.of.MinAwake.MinAsleep Distance..Mi.
## 1 6.625749                   0.9401709          3.34
## 2 6.681416                   0.9384058          3.39
## 3 5.629723                   0.9504132          3.97
## 4 6.293103                   0.9456067          3.48
## 6 5.602871                   0.9270386          4.18
## 7 6.933535                   0.9256018          3.31
head(Afternoon)
##        Rate Ratio.of.MinAwake.MinAsleep Distance..Mi.
## 5  6.470721                   0.9382423          4.44
## 14 7.549148                   0.8788396          7.63
## 20 5.496350                   0.8661258          1.37
## 22 6.017370                   0.8760000          4.03
## 29 5.638086                   0.9490835          4.91
## 30 6.729798                   0.9321267          3.96
boxplot(Morning$Rate,Afternoon$Rate, main = "Comparing Pace of Running (AM vs PM)", names = c("Morning","Afternoon"), ylab = "Run Pace (min/mile)",col=c("thistle","wheat"))

Weekday<-Activity[Activity$Weekday == "Y", c("Rate","Ratio.of.MinAwake.MinAsleep","Distance..Mi.")]
Weekend<-Activity[Activity$Weekday == "N", c("Rate","Ratio.of.MinAwake.MinAsleep","Distance..Mi.")]

boxplot(Weekday$Rate,Weekend$Rate, main = "Comparing Pace of Running (Weekday vs Weekend)", names = c("Weekday","Weekend"),ylab = "Run Pace (min/mile)",col=c("thistle","wheat"))