Loading and preprocessing the data

Unzip and load data

unzip(zipfile = 'activity.zip')
activ_data <- read.csv('activity.csv')

Process and transform data

#format date var
activ_data$date <- as.POSIXct(activ_data$date, format = '%Y-%m-%d')

Create new data frame grouped by total steps per day

total_steps <- group_by(activ_data[complete.cases(activ_data),], date) %>% summarise(total_steps = sum(steps, na.rm = TRUE))

Create histogram of the total number of steps taken each day

ggplot(total_steps, aes(x = total_steps)) +
  geom_histogram(color = 4, fill = "white", bins = 30) + 
  xlab('Total steps per day') +
  ylab('Frequency (count)')

What is mean total number of steps taken per day?

cat(bold('The mean total steps per day is:', mean(total_steps$total_steps)))
## The mean total steps per day is: 10766.1886792453
cat(bold('The median total steps per day is:', median(total_steps$total_steps)))
## The median total steps per day is: 10765

What is the average daily activity pattern?

Create new data set

ts_data <- group_by(activ_data, interval) %>% 
   summarise(mean_steps = mean(steps, na.rm = T))

Plot the data

plot(ts_data$interval, ts_data$mean_steps, type="l", col = "blue", xlab="Interval", ylab="Average Number of Steps",main="Activity by Interval")

Calculate a 5-minute interval that, on average, contains the maximum number of steps

max_interval <- ts_data[which.max(ts_data$mean_steps),1]
max_interval <- unlist(max_interval);
cat("5-min interval with the maximum number of steps:", max_interval)
## 5-min interval with the maximum number of steps: 835

Imputing missing values

Calculate and report the total number of missing values in the dataset

cat('Number of NAs in steps:', sum(is.na(activ_data$steps)))
## Number of NAs in steps: 2304
cat('Number of NAs in date:', sum(is.na(activ_data$date)))
## Number of NAs in date: 0
cat('Number of NA in inteval:', sum(is.na(activ_data$interval)))
## Number of NA in inteval: 0

The strategy used to account for missing data was to impute the mean for any NA.

activ_data$steps[is.na(activ_data$steps)] <- mean(activ_data$steps, na.rm = T)  
cat('Number of NA in steps post imputation:', sum(is.na(activ_data$steps)))
## Number of NA in steps post imputation: 0

Create histogram of the total number of steps taken each day after missing values are imputed

new_total_steps <- group_by(activ_data[complete.cases(activ_data),], date) %>% summarise(total_steps = sum(steps, na.rm = TRUE))

Plot the data

ggplot(new_total_steps, aes(x = total_steps)) +
  geom_histogram(color = 4, fill = "white", bins = 30) + 
  xlab('Total steps per day') +
  ylab('Frequency (count)')

Are there differences in activity patterns between weekdays and weekends?

Create new variable with two levels: weekday and weekend

activ_data$date <- as.Date(activ_data$date)
activ_data$weekday <- weekdays(activ_data$date)
activ_data$weekday = ifelse(activ_data$weekday %in% c('Sunday', 'Saturday'), 'Weekend', 'Weekday')

Create new data frame grouped by total steps per day

new_ts_data <- activ_data %>% group_by(weekday, interval) %>% 
   summarise(mean_steps = mean(steps))
## `summarise()` has grouped output by 'weekday'. You can override using the
## `.groups` argument.

Plot data

xyplot(new_ts_data$mean_steps ~ new_ts_data$interval|new_ts_data$weekday, groups = new_ts_data$weekday, main="Activity by Day of the Week",xlab="5-Minute Interval", ylab="Average Number of Steps",layout=c(1,2), type="l", col=c("red","blue"))

wk_grp_mean <- new_ts_data %>% group_by(new_ts_data$weekday) %>% summarise(across(mean_steps, mean, na.rm = TRUE))

cat(bold('Activity Patterns for Weekends are higher than Weekdays.'))
## Activity Patterns for Weekends are higher than Weekdays.
cat(bold('The mean total steps for weekdays is:', unlist(wk_grp_mean[1,2])))
## The mean total steps for weekdays is: 35.6105811786629
cat(bold('The mean total steps for weekend is:', unlist(wk_grp_mean[2,2])))
## The mean total steps for weekend is: 42.366401336478