This data munging project attempts to clean and process data collected from accelerometers in Samsung Galaxy smartphones. The original data, as well as information about collection and processing, is available at http://archive.ics.uci.edu/ml/datasets/Human+Activity+Recognition+Using+Smartphones.

Once this data has been downloaded, the script run_Analysis.R can be used to process. run_Analysis.R can be run in two ways; as an Rscript via the command line, or by using source() in an interactive session.

If called from the command line run_Analysis.R will write output of the analysis to three files. If the 'UCI HAR Dataset' is not in the current working directory an optional command line argument can be used to specify where to look for the directory. The three output files, which are saved into the current working directory, are:

• mergedData.txt

  This file merges all the data about measurements, individuals, and activities from both the testing and the training set.

• extractedData.txt

  This file processes the merged data and extracts data about individuals, activities, as well as measurement variables with the strings '-mean()-' or '-std()-' in the name.

• tidyData.txt

  This file contains the fully processed data. The extracted data is split by the variables individual and activity, and the mean for each other variable is calculated.

If "sourced" in during an interactive session run_Analysis.R will immediately generate the tidy data, as explained above, and store it in the variable 'tidyDataSet.' There are four functions available in the script:

• mergeData

  Which, as its names suggests, merges all the data. This function has an optional argument 'inDirectory' which can be used if the 'UCI HAR Dataset' is in a different directory.

• extractMeanStd

  This function takes in the merged data set and extracts only the desired columns.

• createTidyData

  Takes the extractedData as an argument and calculates the mean for each variable for every combination of individual and activity.

• run_analysis

  This helper function runs the entire analysis and returns the tidy data set. It is also called when the script is first loaded into R interactively.