PA1_template.md

title	output
Reproducible Research: Peer Assessment 1	html_document keep_md true

Loading and preprocessing the data

Sys.setlocale(category = "LC_ALL", locale = "en_US.utf8")

## [1] "LC_CTYPE=en_US.utf8;LC_NUMERIC=C;LC_TIME=en_US.utf8;LC_COLLATE=en_US.utf8;LC_MONETARY=en_US.utf8;LC_MESSAGES=zh_CN.UTF-8;LC_PAPER=zh_CN.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=zh_CN.UTF-8;LC_IDENTIFICATION=C"

library(dplyr)
library(lattice)
unzip("activity.zip")
activity <- read.csv("activity.csv")
activity_df <- tbl_df(activity)

What is mean total number of steps taken per day?

activity_perday <- summarise(group_by(activity_df, date), steps = sum(steps))
barplot(activity_perday$steps, names.arg = activity_perday$date,
        ylab = "Number of Steps",
        main = "Total Number of Steps Taken per Day"
        )

Mean of above distribution is 1.0766 × 10⁴ while the meadian is 10765.

What is the average daily activity pattern?

activity_interval <- summarise(group_by(activity_df, interval), steps = mean(steps, na.rm = TRUE))
plot(activity_interval$interval, activity_interval$steps,
        type = "l",
        ylab = "Number of Steps",
        xlab = "Time Intervals",
        main = "Average Daily Steps by Time(5 Mins per interval)"
        )

Across all days, the maximum of average number of steps happened on interval 835, which is 206.1698.

Imputing missing values

The total number of rows with NAs is 2304. Use average steps of a given interval to fill the missing value.

activity_filled <- activity_df

for (x in 1:nrow(activity_filled)){
  if(is.na(activity_filled[x,]$steps)){
    mark <- activity_filled[x,]$interval
    activity_filled[x,]$steps <- filter(activity_interval, interval == mark)$steps
  } 
}
rm(x, mark)

Below is what the histgram of the new filled dataset.

activity_filled_perday <- summarise(group_by(activity_filled, date), steps = sum(steps))
barplot(activity_filled_perday$steps, names.arg = activity_filled_perday$date,
        ylab = "Number of Steps",
        main = "Total Number of Steps Taken per Day(Filled)"
        )

Mean of above distribution is 1.0766 × 10⁴ while the meadian is 1.0766 × 10⁴. There should be no big change from original dataset.

Are there differences in activity patterns between weekdays and weekends?

activity_week <- cbind(activity_filled, week = NA)

for (x in 1:nrow(activity_week)){
    activity_week[x,]$week <- ifelse(test = (weekdays(strptime(activity_week[x,]$date, format = "%Y-%m-%d"), TRUE) %in% c("Sat", "Sun")), 
                                     yes = "weekend",
                                     no = "weekday") 
}
rm(x)

activity_week_interval <- summarise(group_by(activity_week, interval, week), steps = mean(steps))
xyplot(steps ~ interval | week, data = activity_week_interval,
       layout = c(1, 2),
       type = "l",
       ylab = "Number of Steps",
       xlab = "Time Intervals",
       main = "Average Daily Steps by Time(5 Mins per interval)"
       )

Above chart suggests there indeed some differences between weekdays and weekend.