Goals 1
Despite the fact that “Time.in.bed” and “TST_mins” are usually used with different meanings in sleep studies, I could confirm that in this data set “Time.in.bed” means “TST_mins” through HW3. So, I will omit the part to check sleep efficiency (C.Goals 1 in HW3).
head(sleep_data$Time.in.bed)
[1] 512 16 510 404 432 438head(sleep_data$TST_mins)
[1] 512 16 510 404 432 438mean() #mean median() #median min() #lowest value max() #highest value sd() #standard deviation var() #variance (standard deviation squared) IQR() #interquartile range
#Presented means, medians, mins, maxes, and SDs for sleep quality, time in bed and heart rate according to states (e.g., coffee, tea, exercising and stress)
sleep_data1<-sleep_data[ -c(1,2,5,6,12) ] #sleep_data1 is created for further analyses (2.2-2.4).
sleep_data1 %>% #mean
group_by(coffee_state,tea_state,working_out_state,stress_state) %>%
summarize_all(mean, na.rm = TRUE)
# A tibble: 14 × 7
# Groups: coffee_state, tea_state, working_out_state [7]
coffee_state tea_state working_out_state stress_state Sleep.quality
<chr> <chr> <chr> <chr> <dbl>
1 No No No No 71.9
2 No No No Yes 3
3 No Yes No No 84
4 No Yes No Yes 93
5 No Yes Yes No 79.9
6 No Yes Yes Yes 90.3
7 Yes No No No 75.8
8 Yes No No Yes 99
9 Yes No Yes No 79.3
10 Yes No Yes Yes 82
11 Yes Yes No No 82.0
12 Yes Yes No Yes 80.6
13 Yes Yes Yes No 76
14 Yes Yes Yes Yes 60.5
# … with 2 more variables: Time.in.bed <dbl>, Heart.rate <dbl>sleep_data1 %>% #median
group_by(coffee_state,tea_state,working_out_state,stress_state) %>%
summarize_all(median, na.rm = TRUE)
# A tibble: 14 × 7
# Groups: coffee_state, tea_state, working_out_state [7]
coffee_state tea_state working_out_state stress_state Sleep.quality
<chr> <chr> <chr> <chr> <dbl>
1 No No No No 91
2 No No No Yes 3
3 No Yes No No 82
4 No Yes No Yes 93
5 No Yes Yes No 78
6 No Yes Yes Yes 92
7 Yes No No No 76
8 Yes No No Yes 99
9 Yes No Yes No 82
10 Yes No Yes Yes 82
11 Yes Yes No No 82
12 Yes Yes No Yes 77
13 Yes Yes Yes No 77.5
14 Yes Yes Yes Yes 60
# … with 2 more variables: Time.in.bed <dbl>, Heart.rate <dbl>sleep_data1 %>% #min
group_by(coffee_state,tea_state,working_out_state,stress_state)%>%
summarize_all(min, na.rm = TRUE)
# A tibble: 14 × 7
# Groups: coffee_state, tea_state, working_out_state [7]
coffee_state tea_state working_out_state stress_state Sleep.quality
<chr> <chr> <chr> <chr> <int>
1 No No No No 16
2 No No No Yes 3
3 No Yes No No 75
4 No Yes No Yes 93
5 No Yes Yes No 59
6 No Yes Yes Yes 86
7 Yes No No No 58
8 Yes No No Yes 98
9 Yes No Yes No 53
10 Yes No Yes Yes 82
11 Yes Yes No No 64
12 Yes Yes No Yes 76
13 Yes Yes Yes No 50
14 Yes Yes Yes Yes 54
# … with 2 more variables: Time.in.bed <int>, Heart.rate <int>sleep_data1 %>% #max
group_by(coffee_state,tea_state,working_out_state,stress_state)%>%
summarize_all(max, na.rm = TRUE)
# A tibble: 14 × 7
# Groups: coffee_state, tea_state, working_out_state [7]
coffee_state tea_state working_out_state stress_state Sleep.quality
<chr> <chr> <chr> <chr> <int>
1 No No No No 100
2 No No No Yes 3
3 No Yes No No 100
4 No Yes No Yes 93
5 No Yes Yes No 100
6 No Yes Yes Yes 93
7 Yes No No No 90
8 Yes No No Yes 100
9 Yes No Yes No 97
10 Yes No Yes Yes 82
11 Yes Yes No No 100
12 Yes Yes No Yes 93
13 Yes Yes Yes No 93
14 Yes Yes Yes Yes 68
# … with 2 more variables: Time.in.bed <int>, Heart.rate <int> sleep_data1 %>% #sd
group_by(coffee_state,tea_state,working_out_state,stress_state)%>%
summarize_all(sd, na.rm = TRUE)
# A tibble: 14 × 7
# Groups: coffee_state, tea_state, working_out_state [7]
coffee_state tea_state working_out_state stress_state Sleep.quality
<chr> <chr> <chr> <chr> <dbl>
1 No No No No 36.3
2 No No No Yes NA
3 No Yes No No 8.12
4 No Yes No Yes NA
5 No Yes Yes No 10.9
6 No Yes Yes Yes 3.79
7 Yes No No No 10.1
8 Yes No No Yes 1.41
9 Yes No Yes No 14.4
10 Yes No Yes Yes NA
11 Yes Yes No No 9.19
12 Yes Yes No Yes 6.05
13 Yes Yes Yes No 9.53
14 Yes Yes Yes Yes 5.80
# … with 2 more variables: Time.in.bed <dbl>, Heart.rate <dbl>ggplot(sleep_data1, aes(coffee_state)) + geom_bar() + theme_bw() + labs(title = "Conditions before sleep", y = "Number of Responses", x = "Drank coffee")
ggplot(sleep_data1, aes(tea_state)) + geom_bar() + theme_bw() + labs(title = "Conditions before sleep", y = "Number of Responses", x = "Drank tea")
ggplot(sleep_data1, aes(working_out_state)) + geom_bar()+ theme_bw() + labs(title = "Conditions before sleep", y = "Number of Responses", x = "Working-out state")
ggplot(sleep_data1, aes(stress_state)) + geom_bar()+ theme_bw() + labs(title = "Conditions before sleep", y = "Number of Responses", x = "Stress state")
Goals 2
Goals 2.2-2.4 (C.Goals 2.2-2.4 which I described in HW3).
2.1 There is a paper that "worse mood" was reported from the subjects who woke up early compared to those who fell asleep late assuming they had same amount of sleep. So, I will see the realtionship between **time to fall asleep or time to wake up vs. sleep quality**.
2.2 I will see the effects of coffee consumed during the day on sleep features -> **Drank coffee vs. time to fall asleep, time to wake up, sleep quality or TST**
2.3 I will see the effects of tea consumed during the day on sleep features -> **Drank tea vs. time to fall asleep, time to wake up, sleep quality or TST**
2.4 I will see the effects of stress during the day on sleep features -> **Stressful day vs. time to fall asleep, time to wake up, sleep quality or TST**
2.5 I will see the effects of excercise during the day on sleep features -> **Worked out vs. time to fall asleep, time to wake up, sleep quality or TST**#2.1 There is a paper that higher portion of "worse mood" was reported from the subjects who woke up early compared to those who fell asleep late assuming they had same amount of sleep. So, I will see the relationship between time to fall asleep or time to wake up vs. sleep quality and mood at awake**. In addition, others of sleep features will be analyzed.
#2.1-1 relationship between time to go to the bed and mood at awake
ggplot(sleep_data, aes(Start,Mood.at.awake)) + geom_point()
#2.1-2 relationship between time to wakeupand mood at awake
ggplot(sleep_data, aes(End, Mood.at.awake)) + geom_point()
#No relationship was observed between time to sleep or time to wake up and mood at awake. Even though there was a limitation that the sleep amount was not the same between early sleepers and late sleepers, I wanted to check the time to go to the bed still may affect the mood at awake. But, I couldn't see any relationship between them. Furthermore, the relationship of time to wake up to the mood at awake was observed, whcih still didn't show any correlation either.
#2.1-3 relationship between time to go to the bed and sleep quality
ggplot(sleep_data, aes(Start, Sleep.quality)) + geom_point()
#2.1-4 relationship between each of time to wake up sleep quality
ggplot(sleep_data, aes(End, Sleep.quality)) + geom_point()
#Regardless of time to sleep or wake up, sleep quality is generally high over the subjects.
#2.2 I will see the effects of coffee consumed during the day on sleep features -> **Drank coffee vs. time to fall asleep, time to wake up, sleep quality or TST**
ggplot(sleep_data, aes(Start,coffee_state)) + geom_point()
ggplot(sleep_data, aes(End,coffee_state)) + geom_point()
#There seems to be no effect of drinking coffee before sleep on the time to sleep or to wake up.
ggplot(sleep_data1, aes(coffee_state, Sleep.quality)) + geom_violin()
ggplot(sleep_data1, aes(coffee_state, Time.in.bed)) + geom_violin()
#For those who drank coffee showed better sleep quality and higher amount of time in bed compared to the group who had no coffee.
#2.3 I will see the effects of tea consumed during the day on sleep features -> **Drank tea vs. time to fall asleep, time to wake up, #sleep quality or TST**
ggplot(sleep_data, aes(Start,tea_state)) + geom_point()
ggplot(sleep_data, aes(End,tea_state)) + geom_point()
#Drinking tea did not show the effects of time to sleep or to wake up either.
ggplot(sleep_data1, aes(tea_state, Sleep.quality)) + geom_violin()
ggplot(sleep_data1, aes(tea_state, Time.in.bed)) + geom_violin()
#Drinking tea showed better sleep quality and higher amount of time to sleep compared to no drinking group since those who had no tea who had overall results of sleep amount and quality while all of the subjects who had tea showed better sleep quality and higher amount of sleep.
#2.4 I will see the effects of stress during the day on sleep features -> **Stressful day vs. time to fall asleep, time to wake up, sleep quality or TST**
ggplot(sleep_data, aes(Start,stress_state)) + geom_point()
ggplot(sleep_data, aes(End,stress_state)) + geom_point()
#Stress did not show the effects of time to sleep or to wake up either.
ggplot(sleep_data1, aes(stress_state, Sleep.quality)) + geom_violin()
ggplot(sleep_data1, aes(stress_state, Time.in.bed)) + geom_violin()
#Compared to those who had stress, most of people who had no stress showed higher amount of time in bed.
#2.4 I will see the effects of excercise during the day on sleep features -> **Worked out vs. time to fall asleep, time to wake up, sleep quality or TST**
ggplot(sleep_data, aes(Start,working_out_state)) + geom_point()
ggplot(sleep_data, aes(End,working_out_state)) + geom_point()
#Working-out state did not show the effects of time to sleep or to wake up either.
ggplot(sleep_data1, aes(working_out_state, Sleep.quality)) + geom_violin()
ggplot(sleep_data1, aes(working_out_state, Time.in.bed)) + geom_violin()
#Compared to those who didn't work out, most of people who worked out showed higher amount of time in bed and higher quality of sleep as well.
#For fun, extra analyses were observed.
#relationship between time in bed(TST) and sleep quality
ggplot(data = sleep_data1, mapping = aes(x = Time.in.bed, y = Sleep.quality)) +
geom_point(mapping = aes(color = Sleep.quality)) +
geom_smooth()
#This figure shows that sleep quality has a positive relationship with the amount of total sleep.
#relationship between time in bed(TST) and sleep quality
ggplot(data = sleep_data1, mapping = aes(x = Heart.rate, y = Sleep.quality)) +
geom_point(mapping = aes(color = Sleep.quality)) +
geom_smooth()
#This figure shows that sleep quality has a negative relationship with heart rate.
HW 4.4 For the final project, I will think about ways to show the features of sleep each other in one graph so that all features look cleaner in one graph. If possible, I will apply colors to the graphs as well. For limitations of 2.1, I didn’t control the amount of sleep for early sleepers and late sleepers. So, I will think about a way to make the amount of sleep the same and compare them (early vs. late) in the same condition (the same amount of sleep) to see the effects of time to fall asleep and to wake up on total sleep time and sleep quality.
