PA1_template.md

title	Reproducible Research: Peer Assessment 1
output	html_document keep_md true

Loading and preprocessing the data

The number of steps per 5 minute interval over two months is provided in the file activity.csv.

library(dplyr)
library(lubridate)
stepsDf <- read.csv('activity.csv', stringsAsFactors = FALSE)
head(stepsDf)

##   steps       date interval
## 1    NA 2012-10-01        0
## 2    NA 2012-10-01        5
## 3    NA 2012-10-01       10
## 4    NA 2012-10-01       15
## 5    NA 2012-10-01       20
## 6    NA 2012-10-01       25

steps <- tbl_df(stepsDf)
stepsByDay <- 	steps %>%  
				group_by(date) %>% 
				summarise(stepTotal = sum(steps, na.rm = TRUE))
hist(	stepsByDay$stepTotal, 
		breaks = 10, 
		main = "Histogram of Total Number of Steps Taken Daily",
		xlab = "Total Steps")

What is mean total number of steps taken per day?

mean(stepsByDay$stepTotal, na.rm = TRUE)

## [1] 9354.23

What is median total number of steps taken per day?

median(stepsByDay$stepTotal, na.rm = TRUE)

## [1] 10395

What is the average daily activity pattern?

stepsByInterval <- 	steps %>% 
					group_by(interval) %>%
					summarise(stepsMean = mean(steps, na.rm=TRUE)) %>%
					mutate(	t = sprintf('%04d', interval), 
							interval_t = paste(substr(t, 1, 2), substr(t, 3, 4), '00', sep = ':' ),
							tod = hms(interval_t))
with(stepsByInterval, plot(	x = tod, 
							y = stepsMean, 
							xaxt = 'n', 
							type = 'l', 
							main = 'Average Daily Activity',
							ylab = 'Average Steps',
							xlab = 'Time of Day'
							))
axis(side = 1, at = seq(0, 86400, 3600), labels = paste0(seq(0, 24, 1), ':00'))

What is the interval that, on average, contains the maximum number of steps

stepsByInterval[[which(stepsByInterval$stepsMean == max(stepsByInterval$stepsMean)), 'interval']]

## [1] 835

Imputing missing values

How many missing values are there?

sum(is.na(stepsDf[,1]))

## [1] 2304

Imputing strategy

The missing values in the original dataset are replaced with the mean of the same interval across those observations that were not missing

fullSteps <- 	merge(	steps, stepsByInterval,
						by.x='interval', by.y='interval') %>% 
				mutate(newsteps = ifelse(is.na(steps), stepsMean, steps))
newSteps <- fullSteps %>%
			select(steps = newsteps, date, interval) %>% 
			arrange(date, interval)
head(newSteps)

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

What does it look like now?

stepsByDay <- 	newSteps %>%  
				group_by(date) %>% 
				summarise(stepTotal = sum(steps, na.rm = TRUE))
hist(	stepsByDay$stepTotal, 
		breaks = 10, 
		main = "Histogram of Total Number of Steps Taken Daily (Missing Values Imputed)",
		xlab = "Total Steps")

What is mean total number of steps taken per day?

mean(stepsByDay$stepTotal, na.rm = TRUE)

## [1] 10766.19

What is median total number of steps taken per day?

median(stepsByDay$stepTotal, na.rm = TRUE)

## [1] 10766.19

While it appears unusual that the mean and median are identical, there are a number of days that the steps value for every interval was missing. This imputing strategy resulted in those days having the mean sum of steps. How many days does this affect?

sum(stepsByDay[, 'stepTotal'] == mean(stepsByDay$stepTotal, na.rm = TRUE))

## [1] 8

Are there differences in activity patterns between weekdays and weekends?

Create a factor with 2 values - week and weekend

stepdays <- newSteps %>% 
				mutate(dow = weekdays(as.Date(date))) %>%
				mutate(dow = factor(dow, 
									levels = c(	"Monday", "Tuesday", "Wednesday", 
												"Thursday", "Friday", "Saturday", "Sunday"), 
									labels = c(	"week", "week", "week", "week", 
												"week", "weekend", "weekend"))) %>%
				group_by(interval, dow) %>%
				summarise(stepsMean = mean(steps, na.rm = TRUE)) 
stepsAsDf = as.data.frame(stepdays)
stepsAsDf <- 	stepsAsDf %>%	
				mutate(	t = sprintf('%04d', interval), 
						interval_t = paste(substr(t, 1, 2), substr(t, 3, 4), '00', sep = ':' ),
						tod = hms(interval_t))
par(mfrow = c(2,1), mar = c(0, 0, 0, 0), oma = c(4, 4, 4, 0.5))

with(stepsAsDf[which(stepsAsDf[,2] == 'week'),],
				plot(	x = tod, 
						y = stepsMean, 
						type = 'l', 
						xaxt = 'n',
						ylab = 'Average Steps',
						xlab = 'Time of Day'))
mtext('week', side = 3, line = -1, adj = 0.1)
#axis(side = 1, at = seq(0, 86400, 3600), labels = paste0(seq(0, 24, 1), ':00'))
with(stepsAsDf[which(stepsAsDf[,2] == 'weekend'),],
				plot(	x = tod, 
						y = stepsMean, 
						type = 'l', 
						xaxt = 'n',
						ylab = 'Average Steps',
						xlab = 'Time of Day'))
mtext('weekend', side = 3, line = -1, adj = 0.1)

axis(side = 1, at = seq(0, 86400, 3600), labels = paste0(seq(0, 24, 1), ':00'))
mtext("Time of Day", side = 1, outer = TRUE, line = 2.2)		
mtext("Steps Mean", side = 2, outer = TRUE, line = 2.2)		
mtext('Average Daily Activity', side = 3, outer = TRUE, line = 2.2)