The run_analysis.R script works by first reading the data from the Samsung data which should have been made placed into your working directory. it can be downloaded from here https://d396qusza40orc.cloudfront.net/getdata%2Fprojectfiles%2FUCI%20HAR%20Dataset.zip.Please ensure the file is unzipped and all content placed in your work directory before continuing.
The data to be loaded are in two folders, train and test. The files to be loaded are saved as X_train, X_test, y_train, y_test, subject_train and subject_test. The read.table script is used to read the data from these files into R.
Following this, the row bind(rbind) function is used to merge the relating data sets together by rows. Thus the creation of X_merged (X_train + X_test), y_merged(y_train + y_test) and subject_merged(subject_train,subject_test) dataframes.
The feature.txt file containing the various functions now comes into play.
It is first given more meaningful column names(feature_id, feature_name) with the names() function, following which, the feature_name column is passed to the grep() function, to extract all the mean and standard deviation(std) functions, and this is stored as grep_features. The grep function pulls all features with mean() and std() in their name.
This is used to filter the previously created X-Merged, leaving only the matched values(means and std). The columns names in X-merged are then made more readable by replacing them with the corresponding feature names.
The activity_labels text is then read in, and given more friendly column names. It is then passed to y_merged, following which the y_merge and subject merged are given meaningful column names
The new tidy data set is created from a column bind(cbind) done on (subject_merged, y_merged, X_merged) dataframes. Finally, the aggregate function is used to provide the average(mean) of each variable for each activity and each subject.