PA1_template.Rmd

---
title: "Reproducible Research: Peer Assessment 1"
output: 
  html_document:
    keep_md: true
---


## Loading and preprocessing the data
```{r reading-data}
activity <- read.csv("activity/activity.csv")
summary(activity)
```


## What is mean total number of steps taken per day?
```{r histogram-total-steps}
totalSteps <- with(activity, aggregate(steps, list(date = date), sum))
hist(totalSteps$x, xlab = "Total number of steps taken each day",
     main = "Histogram of total number of steps")
```

The mean total number of steps is
```{r}
mean(totalSteps$x, na.rm = T)
```
and the median number of steps is
```{r}
median(totalSteps$x, na.rm = T)
```


## What is the average daily activity pattern?
```{r average-daily-activity}
avgStepsPerInterval <- with(activity,
                            aggregate(steps, list(interval = interval),
                                      function(x) mean(x, na.rm = T)))
with(avgStepsPerInterval, plot(interval, x, type = "l", ylab = "", xlab = ""))
title(main = "Average number of steps taken, averaged across all days",
      ylab = "Average number of steps taken",
      xlab = "Interval")
```

The interval that contains maximum number of steps is
```{r}
maxIndex <- which.max(avgStepsPerInterval$x)
avgStepsPerInterval$interval[maxIndex]
```


## Imputing missing values
The number of rows containing missing values is
```{r}
sum(!complete.cases(activity))
```

The missing values are replaced by the mean number of steps for the interval
across all days (or 0).
```{r}
cleanedActivity <- activity
cleanedActivity$steps <- as.numeric(apply(activity, 1, function(row) {
        x <- row['steps']
        interval <- row['interval']
        if(is.na(x)) {
                index <- which(
                        avgStepsPerInterval$interval == as.numeric(interval))
                if(length(index) > 0) {
                        x <- avgStepsPerInterval$x[index[1]]
                } else {
                        x <- 0
                }
        }
        x
}))

summary(cleanedActivity)
```

Let's save the cleaned data set.
```{r}
write.csv(cleanedActivity, file = "activity/cleaned_activity.csv",
          row.names = F)
```

Cleaning the data changed a little the histogram of the total number of steps,
because now there are more *average* values.
```{r histogram-cleaned-total-steps}
totalStepsCleaned <- with(cleanedActivity,
                          aggregate(steps, list(date = date), sum))
hist(totalStepsCleaned$x, xlab = "Total number of steps taken each day",
     main = "Histogram of total number of steps")
```

The mean total number of steps in the cleaned data set is
```{r}
mean(totalStepsCleaned$x)
```

and the median total number of steps in the cleaned data set is
```{r}
median(totalStepsCleaned$x)
```


## Are there differences in activity patterns between weekdays and weekends?
Determine which day is a weekday or a weekend day.
```{r}
cleanedActivity$day <- as.factor(apply(cleanedActivity, 1, function(row) {
        date <- as.POSIXct(as.character(row['date']), format = "%Y-%m-%d")
        weekdayNr <- as.integer(strftime(date, "%u"))
        weekday <- "weekday"
        if(weekdayNr == 6 | weekdayNr == 7) {
                weekday <- "weekend"
        }
        weekday
}))
```

Compute mean of steps taken per each interval and each day type.
```{r}
avgStepsPerIntervalAndWeekday <-
        with(cleanedActivity, aggregate(steps, list(interval = interval,
                                                    day = day),
                                        function(x) mean(x, na.rm = T)))

cleanedActivity$stepsAvg <- apply(cleanedActivity, 1, function(row) {
        interval <- as.integer(row['interval'])
        day <- row['day']
        index <- which(avgStepsPerIntervalAndWeekday$interval == interval &
                               avgStepsPerIntervalAndWeekday$day == day)
        avgStepsPerIntervalAndWeekday$x[index]
})
```

```{r avg-steps-interval-weekday}
library(lattice)
plotData <- unique(cleanedActivity[,c(3,4,5)])
xyplot(stepsAvg ~ interval | day, data = plotData, type = "l",
       layout = c(1, 2), xlab = "Interval", ylab = "Number of steps",
       main = "Average number of steps taken")
```