這是小雷在Google Data Analytic課程中最後一階段的實作項目，隨著實作內容開始增多，現在回頭來看這份分析文件，坦白說會覺得缺少了很多，雖然很不想放上來被恥笑，但這總是一個學習的過程，總是由不會到會，由不熟悉到熟悉，也作為自己每個階段回顧時是否有更進一步的證明，這份分析報告中，都是自己從無到有，有遇到很多奇奇怪怪困難的點，一一的查找資料後慢慢做出來，也找了很多看了很多對於健身的相關文章，剛開始做出來時真的是非常難以言喻的興奮，未來小雷可能會再做一份新的研究報告，但是舊的這份會一直留存，是開始另一個階段自我成長的足跡。
這份分析報告有很多地方是需要調整，所以在語法上就不加多做說明，代碼執行結果有很多會跟背景同色，所以就刪除掉，有興趣的朋友可以自行參考或上小雷的Kaggle閱覽~
小雷的Kaggle : 請點我
In [1]:
library(tidyverse)
library(dplyr)
library(tidyr)
library(here)
library(skimr)
library(janitor)
library(lubridate)In [2]:
activity <- read.csv('../input/fitbit/Fitabase Data 4.12.16-5.12.16/dailyActivity_merged.csv')# 活動合併數據
mets <- read.csv('../input/fitbit/Fitabase Data 4.12.16-5.12.16/minuteMETsNarrow_merged.csv')# 分鐘代謝當量
rate <- read.csv('../input/fitbit/Fitabase Data 4.12.16-5.12.16/heartrate_seconds_merged.csv')# 心跳頻率
bmi <- read.csv('../input/fitbit/Fitabase Data 4.12.16-5.12.16/weightLogInfo_merged.csv')# 權重日誌訊息
sleep <- read.csv('../input/fitbit/Fitabase Data 4.12.16-5.12.16/sleepDay_merged.csv')# 每日睡眠
In [3]:
unique(activity$Id)# 33個
unique(mets$Id)# 33個
unique(rate$Id)# 14個
unique(bmi$Id)# 8個
unique(sleep$Id)# 24個In [4]:
str(activity)
str(mets)
str(rate)
str(bmi)
str(sleep)In [5]:
sum(is.na(mets))
sum(is.na(activity))
sum(is.na(rate))
sum(is.na(bmi)) #TURE >> 65na
sum(is.na(sleep))In [6]:
bmi <-
mutate(bmi, height_m = sqrt(WeightKg/BMI)) %>%
mutate(bmi, height_m = round(height_m, digits = 2)) %>%
mutate(bmi, BMI = round(BMI, digits = 1))
bmi <-
rename(bmi, date = Date) %>%
separate(date, into = c("date", "Time"), sep = " ") %>%
clean_names()
bmi <-
select(bmi, id, date, weight_kg, bmi, height_m)In [7]:
activity <-
rename(activity, date = ActivityDate) %>%
separate(date, into = c("date"), sep = " ") %>%
clean_names()
activity <-
select(activity,
-sedentary_active_distance,
-logged_activities_distance)
In [8]:
mets <-
rename(mets, date = ActivityMinute) %>%
separate(date, into = c("date", "time"), sep = " ") %>%
clean_names()
mets <-
mutate(mets, mets = (me_ts / 10))
mets <-
select(mets, -time, -me_ts) %>%
group_by(id, date) %>%
summarise(mets = sum(mets))In [9]:
rate <-
rename(rate, date = Time) %>%
separate(date, into = c("date", "time"), sep = " ") %>%
clean_names()
rate <-
select(rate, -time)
rate <-
group_by(rate, id, date) %>%
summarise_all(list(sum, mean))In [10]:
sleep <-
rename(sleep, date = SleepDay,
total_sleep_hour = TotalMinutesAsleep) %>%
separate(date, into = c("date", "time"), sep = " ") %>%
clean_names()
sleep <- select(sleep, -time)
sleep <-
mutate(sleep, sleep_hour = (total_sleep_hour / 60),
time_in_bed = (total_time_in_bed / 60)) %>%
mutate(sleep, sleep_hour = round(sleep_hour, digits = 2),
time_in_bed = round(time_in_bed, digits =2))In [11]:
activity %>%
summary()In [12]:
mets %>%
summary()In [13]:
rate %>%
summary()In [14]:
bmi %>%
summary()In [15]:
sleep %>%
summary()In [16]:
library(ggplot2)
bmi_sleep <- merge(bmi, sleep, by = "id")
In [17]:
ggplot(data = sleep, mapping = aes(x = date, y = sleep_hour )) +
geom_point(aes(color = date)) +
theme(axis.text.x = element_text(angle = 60)) +
xlab("Date") +
ylab("Sleep Hour") +
labs(title = "Sleep & Day",
subtitle = "Daily Sleep Time Distribution",
fill = "Date")
In [18]:
ggplot(data = activity) +
geom_point(mapping = aes(x = total_steps,
y = calories,
color = date)) +
geom_smooth(aes(x = total_steps,
y = calories)) +
xlab("Total Steps") +
ylab("Calories") +
labs(title = "Step & Calories",
subtitle = "Do More Steps Burn More Calories?",
fill = "Date")
In [19]:
ggplot(data = activity) +
geom_point(mapping = aes(x = lightly_active_minutes,
y = very_active_minutes,
color = calories,
size = calories)) +
geom_smooth((aes(x = lightly_active_minutes,
y = very_active_minutes))) +
xlab("Lightly Active Min") +
ylab("Hightly Active Min") +
labs(title = "Lightly & Heightly Active",
subtitle = "Differences In Calorie Consumption")
In [20]:
ggplot(data = bmi_sleep) +
geom_boxplot(mapping = aes(x = as.factor(bmi),
y = sleep_hour,
fill = as.factor(bmi))) +
theme(axis.text.x = element_text(angle = 0)) +
xlab("BMI") +
ylab("Sleep Hour") +
labs(title = "sleep time",
subtitle = "sleep time",
fill = "BMI")
