PA1_template.md

Reproducible Research: Peer Assessment 1

Loading and preprocessing the data

Read the data assuming that the unzipped csv file is in the working directory

dt <- read.csv("activity.csv", header=T, stringsAsFactors=F)
library(ggplot2)
library(dplyr)

## 
## Attaching package: 'dplyr'
## 
## The following object is masked from 'package:stats':
## 
##     filter
## 
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

What is mean total number of steps taken per day?

# caculate total number of steps by date
sum <- summarise(group_by(dt, date), steps=sum(steps))
sum

## Source: local data frame [61 x 2]
## 
##          date steps
## 1  2012-10-01    NA
## 2  2012-10-02   126
## 3  2012-10-03 11352
## 4  2012-10-04 12116
## 5  2012-10-05 13294
## 6  2012-10-06 15420
## 7  2012-10-07 11015
## 8  2012-10-08    NA
## 9  2012-10-09 12811
## 10 2012-10-10  9900
## ..        ...   ...

# histogram
ggplot(sum, aes(x=steps)) + geom_histogram()

## stat_bin: binwidth defaulted to range/30. Use 'binwidth = x' to adjust this.

# mean and median
summary(sum$steps)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
##      41    8841   10760   10770   13290   21190       8

What is the average daily activity pattern?

# average steps per interval
int <- summarise(group_by(dt, interval), steps=mean(steps, na.rm=T))
# convert integer interval to time format interval
# example: "105"" -> "0105""
int$interval <- sprintf("%04d", int$interval)
# plot
#ggplot(int, aes(x=strptime(int$interval, format="%H%M"), y=steps)) + geom_line()
plot(strptime(int$interval, format="%H%M"), int$steps, type="l", xlab="interval", ylab="average steps")

# which interval contains maximum num of steps
filter(int, steps==max(int$steps))

## Source: local data frame [1 x 2]
## 
##   interval    steps
## 1     0835 206.1698

Imputing missing values

# number of missing value
nrow(dt[!complete.cases(dt), ])

## [1] 2304

# filling in missing value
filled.dt <- dt
# use mean of steps per interval for missing value
filled.dt[!complete.cases(filled.dt), ]$steps <- int$steps
# histogram total number of steps taken each day
filled.sum <- summarise(group_by(filled.dt, date), steps=sum(steps))
ggplot(filled.sum, aes(x=steps)) + geom_histogram()

## stat_bin: binwidth defaulted to range/30. Use 'binwidth = x' to adjust this.

# mean and median
summary(filled.sum$steps)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##      41    9819   10770   10770   12810   21190

There you can see the median is now equal to the mean, which is the result of filling in missing value

Are there differences in activity patterns between weekdays and weekends?

# convert string date to Date object
filled.dt$date <- as.Date(filled.dt$date)
# add another column representing weekday or weekend
weekdays <- weekdays(filled.dt$date)
isWeekend <- factor(weekdays == "Saturday" | weekdays == "Sunday")
levels(isWeekend) <- c("weekday", "weekend")
filled.dt <- mutate(filled.dt, day=isWeekend)
head(filled.dt)

##       steps       date interval     day
## 1 1.7169811 2012-10-01        0 weekday
## 2 0.3396226 2012-10-01        5 weekday
## 3 0.1320755 2012-10-01       10 weekday
## 4 0.1509434 2012-10-01       15 weekday
## 5 0.0754717 2012-10-01       20 weekday
## 6 2.0943396 2012-10-01       25 weekday

# multiplot
weekday <- filled.dt %>% filter(day=="weekday") %>% group_by(interval) %>% summarise(steps=mean(steps)) %>% mutate(interval=sprintf("%04d", interval))
weekend <- filled.dt %>% filter(day=="weekend") %>% group_by(interval) %>% summarise(steps=mean(steps)) %>% mutate(interval=sprintf("%04d", interval))
par(mfrow=c(1,1))
plot(strptime(weekday$interval, format="%H%M"), weekday$steps, type="l", xlab="Interval", ylab="Number of steps", main="Weekday")

plot(strptime(weekend$interval, format="%H%M"), weekend$steps, type="l", xlab="Interval", ylab="Number of steps", main="Weekend")