| title | author | date | output |
|---|---|---|---|
Reproducible Research - Course Project 1 |
Gabriel Verta |
March 6, 2016 |
html_document |
To parse dates we're going to use lubridate, dplyr to parse data, and to plot some graphs ggplot2
library(dplyr)
library(lubridate)
library(ggplot2)
Reading the data, and parsing dates. Grouping activities per day
activities <- read.csv("activity.csv")
activities$date <- ymd(activities$date)
activities_per_day <- group_by(activities, date)
activities_per_interval <- group_by(activities, interval)
Histogram of the total number of steps taken each day:
items <- tapply(activities$steps, activities$date, FUN=sum)
barplot(items, las=2, cex.names = 0.5)
Mean and median number of steps taken each day. Time series plot of the average number of steps taken.
mean_and_median <- summarise(activities_per_day, mean=mean(steps, na.rm = TRUE), n=median(steps, na.rm = TRUE))
plot(mean_and_median$mean, type="l", col="red", ylab = "Steps", xlab = "Day", las=2, main="Average number of steps taken")
lines(mean_and_median$median, type="l", col="blue")
legend("topright", legend=c("mean", "median"), pch=1, col=c("red", "blue"))
The 5-minute interval that, on average, contains the maximum number of steps
average_of_steps_by_interval <- summarise(activities_per_interval, average=mean(steps, na.rm = TRUE))
ranking_of_intervals <- arrange(average_of_steps_by_interval, desc(average))
print(ranking_of_intervals$interval[1])
Code to describe and show a strategy for imputing missing data
- Total number of missing values in the dataset
nas <- is.na(activities$steps)
total_empty <- sum(nas)
percentage_empty <- mean(nas)
# Total empty results
print(total_empty)
# Percentage
print(paste(sprintf("%.2f", percentage_empty * 100), "%", sep=""))
- Using average of that 5-minute to populate NAs
not_empty <- mutate(activities)
na_intervals <- not_empty$interval[nas]
average_na_intervals <- sapply(na_intervals, function(item) {average_of_steps_by_interval$average[average_of_steps_by_interval$interval==item]})
not_empty$steps[nas] = average_na_intervals
#average_of_steps_by_interval$average[average_of_steps_by_interval$interval==0]
Histogram of the total number of steps taken each day after missing values are imputed
not_empty_items <- tapply(not_empty$steps, not_empty$date, FUN=sum)
par(mfrow=c(2,1))
barplot(items, las=2, cex.names = 0.5, main="Histogram with missing values (NAs)")
barplot(not_empty_items, las=2, cex.names = 0.5, main="Histogram without missing values (NAs)")
As we can see, missing values can change the final result of an analysis.
Panel plot comparing the average number of steps taken per 5-minute interval across weekdays and weekends
activities_with_days <- mutate(activities, weekday=weekdays(date))
activities_by_day_interval <- group_by(activities_with_days, weekday, interval)
results_per_day <- summarise(activities_by_day_interval, steps=mean(steps, na.rm=TRUE))
qplot(interval, steps, data=results_per_day, facets= weekday ~ .) + geom_line()