PA1_template.Rmd

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


## Loading and preprocessing the data
```{r}
activityZip <- 'activity.zip';
activityFilename <- 'activity.csv';
if (!file.exists(activityFilename))
        unzip(activityZip);

activity <- read.csv(activityFilename, header = TRUE, na.strings = "NA");

summary(activity)

```

## What is mean total number of steps taken per day?

Aggregate activity by date - sum steps for each day
```{r}
aggregatedByDate <- aggregate(. ~ date, data=activity, FUN = sum);
```

Histogram of total steps per day
```{r}
hist(aggregatedByDate$steps, breaks = 30, xlab = "Total steps per day", main = "Histogramm of total steps per day", col=3);
```

Calculate mean of total steps per day 
```{r}
meanTotalSteps <- mean(aggregatedByDate$steps);
meanTotalSteps
```

Calculate median of total steps per day 
```{r}
medianTotalSteps <- median(aggregatedByDate$steps)
medianTotalSteps
```

## What is the average daily activity pattern?
Aggregate activity by intervals - average steps for each interval    

```{r}
aggregatedByInterval <- aggregate(steps ~ interval, data=activity, FUN = mean);
```

Plot of average steps per interval
```{r}
plot(x = aggregatedByInterval$interval, y = aggregatedByInterval$steps, type = "l", xlab = "Interval", ylab="Average steps" )
```

Which interval has the maximum number of steps averaged by day
```{r}
aggregatedByInterval[which.max(aggregatedByInterval$steps), ];
```


## Imputing missing values

Total number of missing values in the dataset (i.e. the total number of rows with NAs)
```{r}
sum(!complete.cases(activity))
```

Fill out missing values with average value for interval
```{r}
activityNoNA <- activity;
for (i in 1:nrow(activityNoNA)) {
    if (is.na(activityNoNA$steps[i])) {
        activityNoNA$steps[i] <- aggregatedByInterval[activityNoNA$interval[i] == aggregatedByInterval$interval, ]$steps;
    }
}

summary(activityNoNA)
```

Histogram of the total number of steps taken each day 
```{r}
aggregatedByDateImputed <- aggregate(. ~ date, data=activityNoNA, FUN = sum);
hist(aggregatedByDateImputed$steps, breaks = 30, xlab = "Total steps per day", main = "Histogramm of total steps per day", col=3);

```
 
Imputing effect on estimates 

Mean imputed total steps:
```{r}
meanImputedTotalSteps <- mean(aggregatedByDateImputed$steps)
meanImputedTotalSteps
```

Median imputed total steps:
```{r}
medianImputedTotalSteps <- median(aggregatedByDateImputed$steps)
medianImputedTotalSteps
```

Difference of mean value:
```{r}
meanabsdiff = abs((meanTotalSteps - meanImputedTotalSteps)/meanTotalSteps*100);
meanTotalSteps - meanImputedTotalSteps
```
mean differs in percentage `r meanabsdiff` %  - so mean is not changed


Difference of median value:
```{r}
medianabsdiff = abs((medianTotalSteps - medianImputedTotalSteps)/medianTotalSteps*100);
medianTotalSteps - medianImputedTotalSteps

```
median differs in percentage `r medianabsdiff` %
 
## Are there differences in activity patterns between weekdays and weekends?

Introduce new variable dayType weekend or weekday 
```{r}
activityNoNA['dayType']  <- weekdays(as.Date(activityNoNA$date));
activityNoNA[activityNoNA$dayType %in% c('Saturday','Sunday'), ]$dayType <- "weekend";
activityNoNA[activityNoNA$dayType != "weekend", ]$dayType <- "weekday";
activityNoNA$dayType <- as.factor(activityNoNA$dayType);
summary(activityNoNA$dayType)

```

Average steps in interval by day type across all days and make a plots of steps by day type

```{r}
aggregatedByDateTypeImputed <- aggregate(steps ~ interval + dayType, activityNoNA, mean);

library(ggplot2)

qplot(interval, steps, data = aggregatedByDateTypeImputed, geom=c("line"), xlab = "Interval", ylab = "Number of steps" ) + facet_wrap(~ dayType, ncol = 1)

```