The file run_analysis.R, when sourced, will provide three
interesting data frames and a command to write the tidy version of the
data to disk. Note that this script assumes that the Human Activity
Recognition Using Smartphones Dataset is located in the dataset
directory at the same level as this file, though that dataset is not
included in this repository (it is simply too large and it is readily
accessible online).
gacd.rawData stores the complete raw data from the 'Human Activity
Recognition Using Smartphones Dataset'. That frame has a column for
each reading in the datasets plus columns for userId and activityId.
It has rows for each reading in the test and train datasets.
gacd.data is a cleaned-up subset of that data. The differences are
(a) that this version uses descriptive variable names for the columns,
(b) this version contains an activityName column with a descriptive
activity name, and (c) this only contains measurement columns that
deal with standard deviation or mean.
gacd.tidyData is an even further cleaned version of this data. It
is ordered by userId and activityName, and it has one row per
userId/activityName pair. It has a column for each of the mean and
standard deviation columns in gacd.data, where the value for that
column is the average measurement (of that value) for the
userId/activityName pair.
gacd.writeToTxt gives you a simple wrapper command to write any of
these datasets to a space-delimited file.