ReproducibleResearchProject1.Rmd

---
title: "Reproducible Research: Peer Assessment 1"

output:

html_document:

keep_md: yes
---

## Load the data

```{r}
data <- read.csv("activity.csv")

```

## What is mean total number of steps taken per day?
```{r}
StepsByDay <- aggregate(steps ~ date, data, sum)
hist(StepsByDay$steps, main = paste("Total Number of Steps Each Day"), col="green", xlab="Number of Steps")
StepsMean <- mean(StepsByDay$steps)
StepsMedian <- median(StepsByDay$steps)

```
The mean is `r round(StepsMean)` and the median is `r StepsMedian`

## What is the average daily activity pattern?
```{r}
StepsperInterval <- aggregate(steps ~ interval, data, mean)
plot(StepsperInterval$interval,StepsperInterval$steps, type="l", xlab="Interval", ylab="Number of Steps",main="Average Number of Steps per Day by Interval")
MaxInterval <- StepsperInterval[which.max(StepsperInterval$steps),1]
```
The 5 minute interval that contains the average maximum number of steps is the `r MaxInterval`

## Imputing missing values

There are a number of days/intervals where there are missing values (coded as NA). The presence of missing days may introduce bias into some calculations or summaries of the data.
```{r}
missing <- sum(!complete.cases(data))
```

There are `r missing` items missing in dataset

We will replace the missing items by the average step of that day ith below formula
```{r}
fill.value <- function(steps, interval) {
      filled <- NA
      if (!is.na(steps))
            filled <- c(steps)
      else
            filled <- (StepsperInterval[StepsperInterval$interval==interval, "steps"])
      return(filled)
}
```
And create the new dataset
```{r}
dataUpdated <- data
dataUpdated$steps <- mapply(fill.value, dataUpdated$steps, dataUpdated$interval)

```
This is now the histogram with NA filled in
```{r plot#, eval = TRUE, include = TRUE, fig.path = "figure/", fig.keep = "high", fig.show = "asis"}
StepsByDayUpdated <- aggregate(steps ~ date, dataUpdated, sum)
hist(StepsByDayUpdated$steps, main = paste("Total Number of Steps Each Day"), col="green", xlab="Number of Steps")
#calculate new results and impact
StepsMeanUpdated <- mean(StepsByDayUpdated$steps)
StepsMedianUpdated <- median(StepsByDayUpdated$steps)
DiffStepsMean =(StepsMeanUpdated-StepsMean)
DiffStepsMedian =(StepsMedianUpdated-StepsMedian)
DiffTotal <- sum(StepsByDayUpdated$steps) - sum(StepsByDay$steps)
```
With the update, the mean is now `r round(StepsMeanUpdated)` and the median is `r StepsMedianUpdated`
As we can see, the mean has not changed but the median has increased and got closer to the mean which makes the data more "normally distributed".
By filling the NA we have added `r DiffTotal` steps to the data.

## Are there differences in activity patterns between weekdays and weekends?

```{r plot#, eval = TRUE, include = TRUE, fig.path = "figure/", fig.keep = "high", fig.show = "asis"}
dataUpdated$daytype <- ifelse(weekdays(as.Date(dataUpdated$date)) == "Saturday" | weekdays(as.Date(dataUpdated$date)) == "Sunday", "weekend", "weekday")

StepsperInterval <- aggregate(steps ~ interval + daytype, dataUpdated, mean)

library(lattice)

xyplot(StepsperInterval$steps ~ StepsperInterval$interval|StepsperInterval$daytype, main="Average Steps per Day by Interval",xlab="Interval", ylab="Steps",layout=c(1,2), type="l")
```