acheckas / gcdcourseproject

Getting and Cleaning Data Course Project

This repo contains the course project for Week 4.

Included in the repo are the following assets:

• This ReadMe markdown document you're currently viewing
• A run_analysis.R script
• A Codebook markdown document
• In the output directory, there is a tidy data text file representing the last run of the application

About the run_analysis.R script

This script is responsible for generating the output data provided that the UCI HAR Dataset is available in the working directory, in a folder called "UCI HAR Dataset", when run_analysis.R is run.

The script performs the following actions:

• Load all the relevant data for both the training and test data sets
• Add the activity and subject columns to each set of data
• Set the header names from the independent features vector, adding labels for the two extra columns that were introduced.
• Combine the test and training data sets
• Filter out the columns that don't include standard deviations or means
• Create a summary set from the filtered data that is comprised of the means of the remaining columns and present it in a tidy dataset.
• Write the data out to a file in the output directory.

Requirements

This script requires three non-base libraries to be installed in order to run:

• dplyr
• tidyr
• data.table

Output

The output that's generated is a summary of the mean of the original UCI HAR Dataset's collected means and standard deviations for each measurement provided, grouped by subject and by activity.The data is a tidy summarization of the original data. In that it is clearly labeled, has one measurement in each column, and has one observation of that set of variables in each row which ,in this case, is a mean for each observation for each activity for each subject, and there is only one table and thus one file. The top row of the output file includes all of the variable names for the data set which are as human read-able as the sensor data allows.

Loading the output file

When the script is run, it will generate a file named "acHARSummary.txt" in the output subdirectory of your current working directory. To load the file into your R interpreter, you can run:

outputTable <- read.table("./output/acHARSummary.txt",header=TRUE)

The expected data frame is 180 x 81 with a subject, activity and 178 assorted sensor means.

Codebook.md

The Codebook contains all the modifications and updates to the original codebook along with the data to indicate all the variables and summaries calculated, along with units, and any other relevant information.

Sample Dataset

The UCI HAR Dataset that was used to generate this code is also included in the top level of the repo for convenience of reviewers.

The original dataset can be located at http://archive.ics.uci.edu/ml/datasets/Human+Activity+Recognition+Using+Smartphones#