/GetData_CourseProject

Course Project for Coursera Getting and Cleaning Data 2014

Primary LanguageR

Course Project

Course Project for Coursera Getting and Cleaning Data 2014 course

This repository consists of two other files:

  • run_analysis.R contains the R code which downloads, filters, and tidies up the data
  • uci_har_tidy.txt contains the data output by run_analysis_R after it has been filtered and tidied

Data Source

The raw data comes from the HCI Machine Learning Repository and is derived from recordings of 30 subjects performing activities of daily living (ADL) while carrying a waist-mounted smartphone (Samsung Galaxy S II) with embedded inertial sensors.

Because the raw input data file is very large, it isn't included in the repo. Instead, the run_analysis.R includes intructions that first downloads the data and description file set and extracts the archive.

Data Cleaning and Tidying

The rest of the run_analysis.R script performs the following operations on the files contained in HCI's archive:

  1. To merge the training and test data sets to create one data set:
    1. It reads in the training data set from X_train.txt, which consists of measurements, and appends an activity column consisting of the training labels from the file y_train.txt and a subject.ID column consisting of the ID for the subject (volunteer) from the file subject_train.txt
    2. It does the same thing for the testing data set
    3. It combines the training and testing data set by row
  2. To extract only the measurements on the mean and standard deviation for each signal:

    1. It reads in the names of the features from features.txt, verifying that the features are listed in the same order as the columns in X_train.txt and X_test.txt

    2. It figures out the relevant features by looking for std() or var() in the variable names.

      Note that we only look for means that have been computed generally rather than done for a specific purpose (e.g. the meanFreq feature) and computed by averaging the signals in a signal window sample (the angle() variable collumns).

  3. To give descriptive activity names to the activities in the data set:

    1. It reads in descriptions of the activities

    2. It converts the activities in the data set, coded as a number from 1-6, into a factor with descriptive labels from the file activity_labels.txt. The mapping is as follows:

       1 WALKING
 2 WALKING_UPSTAIRS
 3 WALKING_DOWNSTAIRS
 4 SITTING
 5 STANDING
 6 LAYING
  4. To label the columns in data set with descriptive variable names:
    1. It first corrects the typo "BodyBody" to "Body". We assume it's a typo, based on descriptions in the README.txt file in the archive and the fact that the feature names stay distinct after the correction.
    2. We replace the abbreviations by more descriptive names, except for the standard deviation for which we assign the most common industry-accepted abbreviation "SD".
  5. To create a second, independent tidy data set with the average of each variable for each activity and each subject:
    1. It groups the data by subject ID and by activity
    2. It summarizes the data in the groups by averaging

Data Output

The summarized tidy data set is output to file uci_har_tidy.txt in a simple space-separated text file with the first row serving as a header with the column names.