#run_analysis.R

##Purpose

• Merge, label and summarize the "UCI HAR Dataset"
• The output contains the calculated mean by subject and activity of the means and standard deviations from the raw dataset.
• See CodeBook.md for details on input and output data

##Pre-requisites

• The R code was developed and tested on "R version 3.1.2 (2014-10-31)" on Max OS X 10.9, it has not been tested on any other configuration
• The working directory needs to set to the root of the "UCI HAR Dataset"

##How to run

1. Start R
2. Set working directory to the root of the "UCI HAR Dataset"
3. Source the run_analysis.R script

##Input

• The "UCI HAR Dataset" with the following directory structure
  • activitiy_labels.txt
  • features.txt
  • train/
    • subject_text.txt
    • X_test.txt
    • y_test.txt
  • test/
    • subject_train.txt
    • X_train.txt
    • y_train.txt

##Output

• A file named "tidy_df.txt" containting the tidy dataset will be created in the working directory
• The file format is space-separated with quoted character vectors
• The file has a header row describing the columns

##Logic

1. Load reference data (activity_labels.txt and features.txt)
2. Load and label test data set (subject_text.txt, X_test.txt and y_test.txt)
   • Feature labels based on features.txt
3. Merge test data set into one dataframe
4. Load and label training data set (subject_text.txt, X_test.txt and y_test.txt)
   • Feature labels based on features.txt
5. Merge training data set into one dataframe
6. Merge test and training data sets
7. Label activities based on activity_labels.txt
8. Filter features to keep only mean and SD measurements
   • Keep feature names ending in "mean()"" or "sd()""
9. Create tidy dataframe by calculating mean of all features by subject and activity
10. Prepend feature names with "mean_" to indicate that they are mean values
11. Remove "()" and replace "-" with "_" in feature names to make them valid for R use
12. Store tidy dataframe in "tidy_df.txt" file in working directory