date last modified: 2017-04-07
What is mean total number of steps taken per day?
What is the average daily activity pattern?
Are there differences in activity patterns between weekdays and weekends?
This assignment makes use of data from a personal activity monitoring device. This device collects data at 5 minute intervals through out the day. The data consists of two months of data from an anonymous individual collected during the months of October and November, 2012 and include the number of steps taken in 5 minute intervals each day.
URL = "https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2Factivity.zip"included in the source dataset are:
• steps: Number of steps taking in a 5-minute interval (missing values are coded as NA)
• date: The date on which the measurement was taken in YYYY-MM-DD format
• interval: Identifier for the 5-minute interval in which measurement was taken
included in the final dataset are:
• sourcedData: The data.frame that was read into R
• timeOnly: A variable constructed which converts the interval into a time
Note that this stores the current system time data with this value which is to be
disregarded for the analysis
• day: The day of the week
• time: A variable constructed which joins the observation date with the time
• constructedData: This is a transformed version of the sourcedData data.frame
where all NAs have been replaced with the average value for the interval.
• Q1: This is the data.frame associated with Question 1 output.
• Average.StepsPerDay: A list containing the median and mean in answer to Question 1.
• Q2: This is the data.frame associated with Question 2 output.
• activityPeak: A list containing the lower and upper bounds for the most active
interval in answer to Question 2.
• Q3: This a list of two data.frames weekend and weekday associated with
Question 3 output.
## R version R version 3.3.2 (2016-10-31)
## OS mingw32
## dplyr 0.5.0 was used during for the production of this analysis
## stringr 1.1.0 was used during for the production of this analysis
## ggplot2 2.1.0 was used during for the production of this analysis
keep()is an internal function which keeps a record of the important variables
that the script will retain in the workspace following the running of the code.
cleanWorkspace() is a finalising function which removes the <<- variables that
aren't intended to be retained after the running of the function.
keep<-function(...){
#>DESCRIPTION----
# keeps a record of starting globals
# and appends the desired variables that we wish to keep by the end of the
# script, thereby making all other variables and functions that are made during
# the running of this function as temporary.
if(!exists("GlobalEnvKeep")){
GlobalEnvKeep<<-c(ls(envir = as.environment(globalenv())))
}
GlobalEnvKeep<<-c(GlobalEnvKeep,...)
}
keep()
cleanWorkspace<-function(){
rm(list= ls(envir=as.environment(globalenv()))[!(ls(envir=as.environment(globalenv())) %in% GlobalEnvKeep )],
envir=as.environment(globalenv()))
rm(keep,envir=as.environment(globalenv()))
}sourceData() is used to download the data to the workspace and the working folder.
sourceData<-function(src=URL,expectedfiles="activity.csv",workspaceNames="sourcedData"){
if(!file.exists(expectedfiles)){
message("downloading files")
download.file(src,"data.zip")
unzip("data.zip")
file.remove("data.zip")
}
columnClassPreDefined<-c("numeric","character","character")
## This first line will likely take a few seconds to load
if (!exists(workspaceNames)){sourcedData<<- read.csv(expectedfiles,
colClasses=columnClassPreDefined,
header = TRUE,
stringsAsFactors = FALSE,
na.strings = "NA")
#change the date column to a date type field
sourcedData$date<<-as.Date(sourcedData$date)
#pads the interval to become a 4 digit time
sourcedData$interval<<-stringr::str_pad(sourcedData$interval,width=4,side="left",pad="0")
# combines data and time into a single column called time
sourcedData$time<<-paste(sourcedData$date,sourcedData$interval)
sourcedData$time<<-as.POSIXct(strptime(sourcedData$time,"%Y-%m-%d %H%M"))
sourcedData$timeOnly<<-as.POSIXct(strptime(sourcedData$interval,"%H%M"))
sourcedData$day<-weekdays(sourcedData$time)
## retains the sourcedata variable for the project workspace
keep("sourcedData")
}
# organising data----
sourcedData<<-dplyr::tbl_df(sourcedData)
message("Data loaded successfully into the workspace as 'sourcedData'")
##creates a container to keep useful metadata as a list
metadata<<-list(NULL)
metadata$head<<-(head(sourcedData))
metadata$summary<<-summary(sourcedData)
str(sourcedData)
}The call to the function sourceData() loads the data into R.
sourceData(URL)## Data loaded successfully into the workspace as 'sourcedData'
## 'data.frame': 17568 obs. of 6 variables:
## $ steps : num NA NA NA NA NA NA NA NA NA NA ...
## $ date : Date, format: "2012-10-01" "2012-10-01" ...
## $ interval: chr "0000" "0005" "0010" "0015" ...
## $ time : POSIXct, format: "2012-10-01 00:00:00" "2012-10-01 00:05:00" ...
## $ timeOnly: POSIXct, format: "2017-04-07 00:00:00" "2017-04-07 00:05:00" ...
## $ day : chr "Monday" "Monday" "Monday" "Monday" ...
The data when it was loaded in shows there are 17568 observations
and 3 variables. The interval variable represents a
24 hour time and a new variable has been introduced called time joining the date
and time together into a single POSIXct date time. Another variable called day
has also been included which is needed for question 3. The variables are
called steps, date, interval, time, timeOnly, day.
Running summary() on the loaded data shows there are NA's :2304
The observations were taken between 2012-10-01, 2012-11-30.
Q1<-as.data.frame(summarise(group_by(sourcedData,date),
sum(steps,na.rm = TRUE)
))
names(Q1)<-c("date","steps.Total")
keep("Q1")
hist(as.matrix(Q1$steps.Total[Q1$steps.Total!=0]),
xlab="Total number of steps in a day",
main="Plot 1: Distribution for the number of steps taken in a day",
breaks= 20)
rug(as.matrix(Q1$steps.Total[Q1$steps.Total!=0]),col="grey")
abline(v=mean(Q1$steps.Total[Q1$steps.Total!=0]),col="blue",lty=3,lwd=5)
abline(v=median(Q1$steps.Total[Q1$steps.Total!=0]),col="red",lty=5,lwd=1)
legend("topright",legend=c("mean","median"),col=c("blue","red"),lty=c(3,5),lwd=c(5,1))
The total number of steps per day can be seen in the result of the Q1 data.frame
Plot 1 shows an evenly balanced distribution with the most common number of steps per day is between 10000 and 11000. The plot also shows that the mean and median are close to eachother indicating there is very little skew for this set of observations.
Average.StepsPerDay <- list(NULL)
keep(Average.StepsPerDay)
Average.StepsPerDay$mean <- as.integer(mean(Q1$steps.Total[Q1$steps.Total!=0]))
Average.StepsPerDay$median <- as.integer(median(Q1$steps.Total[Q1$steps.Total!=0]))Excluding the days where no data was collected:
The mean number of steps taken per day are 10766 (to nearest whole number).
The median number of steps taken per day are 10765 (to nearest whole number).
Running summary() on the loaded data shows there are NA's :2304
The NA values occur for whole days so the script will attempt to impute values
based on the average interval.
Q2<-as.data.frame(summarise(group_by(sourcedData,timeOnly,interval),
mean(steps,na.rm = TRUE)))
#create a subset of the data where NAs occur
constructedData<-subset(sourcedData,is.na(sourcedData$steps))
#impute the average of the interval where there is the NA
constructedData$steps<-Q2$`mean(steps, na.rm = TRUE)`
#bind the two dataframes together
constructedData<-dplyr::bind_rows(constructedData,subset(sourcedData,!is.na(sourcedData$steps)))
#reorder by time
constructedData<-dplyr::arrange(constructedData, time)
str(constructedData)## Classes 'tbl_df', 'tbl' and 'data.frame': 17568 obs. of 6 variables:
## $ steps : num 1.717 0.3396 0.1321 0.1509 0.0755 ...
## $ date : Date, format: "2012-10-01" "2012-10-01" ...
## $ interval: chr "0000" "0005" "0010" "0015" ...
## $ time : POSIXct, format: "2012-10-01 00:00:00" "2012-10-01 00:05:00" ...
## $ timeOnly: POSIXct, format: "2017-04-07 00:00:00" "2017-04-07 00:05:00" ...
## $ day : chr "Monday" "Monday" "Monday" "Monday" ...
keep("constructedData")
NAcount<-sum(is.na(constructedData$steps))There were 2304 NAs in the sourcedData data.
There are 0 NAs in the constructedData data.
Q2<-as.data.frame(summarise(group_by(constructedData,timeOnly,interval),
mean(steps,na.rm = TRUE)
))
names(Q2)<-c("time","interval","steps.Mean")
keep("Q2")
with(constructedData,
plot(x=timeOnly,
y=steps,
col=rgb(.5,.5,.5,.1),
pch=20,
ylab="Number of Steps",
xlab="Time of Day",
main = "Plot 2: Activity over a day"))
with(Q2,points(x=time,
y=steps.Mean,col="red",
type="l",
ylim=c(0,max(sourcedData$steps,na.rm=TRUE))
))
legend("topright",
legend=c("average number of steps at time interval"),
col=c("red"),
lty=c(1),
lwd=c(1),
box.col="transparent",
bg = "transparent" )
The average activity across a day for the study period can be seen in Plot 2.
activityPeak<-list(NULL)
keep(activityPeak)
activityPeak$most.lower<-format.Date(Q2$time[Q2$steps.Mean==max(Q2$steps.Mean)],"%T")
activityPeak$most.upper<-format.Date(Q2$time[Q2$steps.Mean==max(Q2$steps.Mean)]+299,"%T")The most active time interval is in the morning at 08:35:00 - 08:39:59
A new factor is made in the column weekdayType within the constructedData data.frame
constructedData$weekdayType<-as.factor(
ifelse(constructedData$day=="Saturday"|constructedData$day=="Sunday",
"weekend","weekday"))
Q3<<-split(constructedData,constructedData$weekdayType)
keep("Q3")
str(Q3)## List of 2
## $ weekday:Classes 'tbl_df', 'tbl' and 'data.frame': 12960 obs. of 7 variables:
## ..$ steps : num [1:12960] 1.717 0.3396 0.1321 0.1509 0.0755 ...
## ..$ date : Date[1:12960], format: "2012-10-01" ...
## ..$ interval : chr [1:12960] "0000" "0005" "0010" "0015" ...
## ..$ time : POSIXct[1:12960], format: "2012-10-01 00:00:00" ...
## ..$ timeOnly : POSIXct[1:12960], format: "2017-04-07 00:00:00" ...
## ..$ day : chr [1:12960] "Monday" "Monday" "Monday" "Monday" ...
## ..$ weekdayType: Factor w/ 2 levels "weekday","weekend": 1 1 1 1 1 1 1 1 1 1 ...
## $ weekend:Classes 'tbl_df', 'tbl' and 'data.frame': 4608 obs. of 7 variables:
## ..$ steps : num [1:4608] 0 0 0 0 0 0 0 0 0 0 ...
## ..$ date : Date[1:4608], format: "2012-10-06" ...
## ..$ interval : chr [1:4608] "0000" "0005" "0010" "0015" ...
## ..$ time : POSIXct[1:4608], format: "2012-10-06 00:00:00" ...
## ..$ timeOnly : POSIXct[1:4608], format: "2017-04-07 00:00:00" ...
## ..$ day : chr [1:4608] "Saturday" "Saturday" "Saturday" "Saturday" ...
## ..$ weekdayType: Factor w/ 2 levels "weekday","weekend": 2 2 2 2 2 2 2 2 2 2 ...
par( mfrow= c(2,1), mar = c(4,4,3,1) )
plot(x=Q3$weekday$timeOnly,y=Q3$weekday$steps,
main="Plot3A - Activity on the weekdays",
xlab="time of day",
ylab="number of steps",
col=rgb(.5,.5,.5,.1),
pch=20
)
#get the average for the weekdays
Q3$avg.Weekday<-as.data.frame(summarise(group_by(Q3$weekday,timeOnly,interval),
mean(steps,na.rm = TRUE)
))
names(Q3$avg.Weekday)<-c("time","interval","steps.Mean")
with(Q3$avg.Weekday,points(x=time,
y=steps.Mean,col="red",
type="l",
ylim=c(0,max(sourcedData$steps,na.rm=TRUE))
))
#plots the points
plot(x=Q3$weekend$timeOnly,y=Q3$weekend$steps,
main="Plot3B - Activity on the weekend days",
xlab="time of day",
ylab="number of steps",
col=rgb(.5,.5,.5,.1),
pch=20)
legend("topleft",
legend=c("average number of steps at time interval"),
col=c("red"),
lty=c(1),
lwd=c(1),
box.col="transparent",
bg = "transparent" )
#get the average for the weekends
Q3$avg.weekend<-as.data.frame(summarise(group_by(Q3$weekend,timeOnly,interval),
mean(steps,na.rm = TRUE)
))
names(Q3$avg.weekend)<-c("time","interval","steps.Mean")
with(Q3$avg.weekend,points(x=time,
y=steps.Mean,col="red",
type="l",
ylim=c(0,max(sourcedData$steps,na.rm=TRUE))
))
legend("topleft",
legend=c("average number of steps at time interval"),
col=c("red"),
lty=c(1),
lwd=c(1),
box.col="transparent",
bg = "transparent" )
Plot 3 separates out the weekly routine into weekend and weekdays. It shows that there is a slightly different pattern for weekends.
It appears that within a 5 minute period around 600 steps is what the subjects natural walking pace, and when they are in a rush they can achieve 800 steps within an interval. The peak values occur around the mid afternoon for on the weekends and there is generally more activity throughout the day compared to weekdays, whereas the morning period around 9:00 am for both weekends and weekdays but weekdays are faster paced around this period. Weekdays also show an ocassional short fast paced spikes around lunch time and at around 15:00. ctivity begins to dies dwo naround 19:00 wheras on weekends it can be up to an hour later.